1

MAC Protocols for Terahertz Communication: AComprehensive Survey

Saim Ghafoor, Noureddine Boujnah, Mubashir Husain Rehmani and Alan Davy

Abstract—Terahertz communication is emerging as a futuretechnology to support Terabits per second link with highlightingfeatures as high throughput and negligible latency. However,the unique features of the Terahertz band such as high pathloss, scattering and reflection pose new challenges and resultsin short communication distance. The antenna directionality,in turn, is required to enhance the communication distanceand to overcome the high path loss. However, these featuresin combine negate the use of traditional Medium access pro-tocols. Therefore novel MAC protocol designs are required tofully exploit their potential benefits including efficient channelaccess, control message exchange, link establishment, mobilitymanagement, and line-of-sight blockage mitigation. An in-depthsurvey of Terahertz MAC protocols is presented in this paper. Thepaper highlights the key features of the Terahertz band whichshould be considered while designing an efficient Terahertz MACprotocol, and the decisions which if taken at Terahertz MAClayer can enhance the network performance. Different Terahertzapplications at macro and nano scales are highlighted with designrequirements for their MAC protocols. The MAC protocol designissues and considerations are highlighted. Further, the existingMAC protocols are also classified based on network topology,channel access mechanisms, and link establishment strategies asTransmitter and Receiver initiated communication. The openchallenges and future research directions on Terahertz MACprotocols are also highlighted.

Index Terms—Terahertz band, Terahertz communication net-work, Terahertz technology, Terahertz physical layer, TerahertzMAC layer, Terahertz Channel model, Terahertz Propagationmodel, Terahertz Antenna, Terahertz Transceivers.

I. INTRODUCTION

The demand for wireless data traffic has increased signifi-cantly since the evolution of Internet and Mobile Technologyand is projected to exceed Petabytes by 2021 [1]. The existingwireless technology although reaching the capacity of wiredtechnology, still it is not meeting the demands of future ultra-high bandwidth communication networks. The spectrum atand below 60 GHz still orders of magnitude below the tar-geted Terabits per second (Tbps) link. The Free-space optical(FSO) which operates at Infrared (IR) frequencies also hasseveral issues that limit the practicality of these systems forpersonal wireless communications [2], [3]. In this perspective,the Terahertz (THz) band from 0.1 to 10 Terahertz has thepotential to provide up to Terabits per second (Tbps) linkspeed to satisfy beyond fifth-generation (5G) communication

Saim Ghafoor, Noureddine Boujnah, and Alan Davy are withEmerging Network Laboratory, Telecommunication Systems &Software Group, Waterford Institute of Technology, Ireland. email:s.ghafoor,bnoureddine,adavy@tssg.org.

Mubashir Husain Rehmani is with Cork Institute of Technology, Ireland.email: mshrehmani@gmail.com.

requirements such as high throughput and low latency [3]–[6]. The Terahertz bands offer much larger bandwidth (up to1 THz) than the existing millimeter-wave (mmWave) systems(up to 10 GHz)) [7]–[9].

While, the technology is rapidly advancing with newtransceiver architectures, materials, antenna design, chan-nel/propagation model, and physical layer techniques, therestill exist several research challenges that need to be addressedbefore achieving the Tbps links. Among these different fieldsof interest, Medium Access Control (MAC) is least exploredarea of research in Terahertz communication networks. Theexisting MAC protocols of traditional networks cannot bedirectly applied, because they do not consider the uniquefeatures of Terahertz band like path and molecular loss,multipath, reflection, and scattering. Therefore, novel andefficient MAC protocols are required which should considerthe features of Terahertz bands and antenna requirements.In this paper a comprehensive survey on Terahertz MACprotocols is presented with classification, design issues, andconsiderations, requirements for different application areas andchallenges. The acronyms used commonly throughout thissurvey are shown in Table I.

A. Terahertz Communication: Related survey articles

Table II highlights and summarise the overall survey paperson Terahertz communication without MAC layer protocols.These survey papers cover different application areas, howevercovers mostly the device, antenna, channel, and physicallayer aspects. These include the nano-communication net-works [10]–[12], Internet of nano-things [13], [14], molecularcommunication network [15], [16], nano-sensor network [17],in-body nanonetworks [16], [18], [19], broadband communi-cation [20], vehicular networks [21], wireless indoor/outdoorcommunications which include the office/data-center or smallcells deployment [22], [23] and Terahertz CommunicationNetworks [5], [6], [23]–[30]. Whereas, Table III highlights thesurvey papers which discuss the MAC layer aspects to someextent and the differences with our work.

In [20], [24], the joint impact of Ultra-Dense Network,Multiple Input Multiple Output (MIMO), mmWave and Tera-hertz communications is discussed for supporting the demandsof mobile broadband services. Particularly, the indoor andoutdoor environments are analyzed for noise levels and signalto interference and noise ratio (SINR). In [21], the challenges,and opportunities are mentioned but only with the perspectiveof Terahertz vehicular networks. It also describes brieflydifferent aspects including transceiver design, MIMO antenna

arX

iv:1

904.

1144

1v2

[cs

.NI]

22

Nov

201

9

2

arrays, channel modeling and estimation, interference man-agement, MAC layer design, and standardization. For NanoCommunication Networks, the survey articles are discussedin [10]–[12], [15], [18], with physical layer aspects, prop-agation models, security, in-body communication, biomedi-cal applications, materials, and antenna design, and channelmodeling. The Internet of nano-things (IoNT) for intercon-necting devices at the nanoscale is discussed in [13], [14],[19]. Although, a brief discussion on architecture, channelmodeling and challenges related to MAC and Network layerare mentioned, but only for a specific scenario of the IoNT.A network architecture for wireless nano-sensor networks isgiven in [17], which discusses briefly the challenges relatedto channel modeling, information encoding and protocols fornano-sensor networks. The molecular communication networksurvey is given in [16] for the body area networks. Theseworks although provides a good survey, represent the architec-ture, models, and challenges related to the specific applicationscenario.

The Terahertz communication is still in its design anddevelopment stage and therefore looking at the advantageslike ultra-high bandwidth and negligible latency, the researchdirections should be identified in a manner to advance thetechnology and enhance the system performance. In [22],the Terahertz recent progress is reviewed but only for thepropagation models, antenna and testbed design, and an imple-mentation roadmap are mentioned. For opportunities beyond5G paradigm, an architecture is discussed with possible appli-cation areas in [5]. A survey related to MIMO is given in [28].Some standardization related work is mentioned in [27].Some other survey papers on the usage of Graphene material,weather impact on Terahertz bands and Terahertz antenna arementioned in [31]–[33]. A guest editorial is also publishedrecently on Terahertz communication [34].

With these related survey articles, there are some articlesthat discuss the Terahertz MAC protocols but not with requireddetail, as shown in Table III. In [35], a survey on MACschemes for mmWave and Terahertz wireless communicationis presented. The MAC protocols overall related to Tera-hertz band communication are not fully covered and mostlyMAC strategies related to mmWave are discussed. Only fewchallenges and design issues are mentioned. Whereas, inthis survey paper, detailed work on existing Terahertz MACprotocols with classifications, band features, design issues andconsiderations, application requirements and challenges arediscussed.

B. Contributions of this survey

The existing work, although covers a detailed survey ondevices, antenna, channel, and physical layer aspects. Onlylimited work is available which summarises the TerahertzMAC protocols. No survey paper discusses in detail theTerahertz applications and their requirements; Terahertz bandfeatures; design issues and considerations; and MAC decisionsand challenges towards the design of an efficient MAC proto-col. In this survey paper, these aspects are discussed in detail.The main contributions of this survey paper are:

Fig. 1: Terahertz gap in the electromagnetic spectrum.

• A comprehensive survey of existing Terahertz MACprotocols is presented.

• Classification of existing Terahertz MAC protocols basedon network scale and topologies, channel access mecha-nisms and transmitter/receiver-initiated communication.

• The unique features of the Terahertz band are highlightedto be considered for Terahertz MAC protocols.

• The design issues are highlighted which should be consid-ered while designing efficient Terahertz MAC protocolswith decisions that should be taken at the MAC layer forperformance enhancements.

• The requirements and design challenges for TerahertzMAC protocols for different application areas are dis-cussed.

• Different challenges and future research directions arealso highlighted.

C. Organization of survey

The paper is organized as, Section I presents the intro-duction and literature review. Section II, presents the back-ground on Terahertz band, technology, and MAC protocols.In Section III, different applications of Terahertz band com-munication are discussed with respect to macro and nanoscalecommunication with their requirements. Section IV, highlightsthe unique features of THz band, issues that needs to beconsidered at Physical and MAC layer with decisions whiledesigning an efficient Terahertz MAC protocols. Section V,mentions the topologies so far focused on Terahertz commu-nication networks. Different channel access mechanisms arediscussed in Section VI. The transmitter and receiver-initiatedcommunication are discussed in Section VII. The challengesand future research directions are discussed in Section VIII.Finally, in Section IX, the survey paper is concluded.

II. BACKGROUND ON TERAHERTZ BAND, TECHNOLOGYAND MAC PROTOCOLS

A. Terahertz bands

The Terahertz can be termed as a unit of frequency (onetrillion cycles per second or 1012 Hz) or electromagnetic waveswithin ITU designated band of frequencies. The Terahertzfrequency range (0.1 - 10 THz) is the last span within thewhole electromagnetic wave spectrum and is more commonlyreferred to as Terahertz Gap. They appear between the Mi-crowave and Infrared bands, as shown in Figure 1. Thewavelength of radiation in the Terahertz band range from 1mm to 0.1 mm (or 100 µm). The bands from 100 GHz to 200

3

TABLE I: Acronym definitions used throughout this survey.

Acronyms Definitions5G Fifth generationACK AcknowledgementAP Access pointBER Bit error rateCA Collision avoidanceCSMA Carrier sensing multiple accessCTS Clear to sendDMDS Distributed Maximum depth schedulingESaware Energy and spectrum awareFSO Free space opticalFTDMA Frequency and time division multiple accessLOS Line of sightLTE-A Long term evolution advancedMAC Medium Access ControlMIMO Multiple input multiple outputmmWave Millimeter waveMRAMAC Multiradio assisted MACNLOS Non line of sightOFDM Orthogonal frequency division multiplexingPAM Pulse amplitude modulationPNC Piconet coordinatorPPM Pulse position modulationQoS Quality of serviceRA Random accessRD Rate divisionRTDs Resonant tunnelling diodeRTR Request to receiverRTS Request to sendSDN Software defined networkSINR Signal to interference and noise ratioTABMAC Terahertz assisted beamforming MACTbps Terabits per secondTC Transmission confirmationTCN Terahertz communication networkTDMA Time division multiple accessTHz TerahertzTLAN Terahertz local area networkTR Transmission requestTS-OOK Time spread On-off KeyingTTS Test to sendTPAN Terahertz personal area networkUL UplinkUTC-PD Uni-travelling carrier photo diodesUV UltravioletWLAN Wireless local area networkWPAN Wireless personal area network

GHz are also referred as sub-Terahertz band [43], as it beginsat a wavelength of one millimeter and proceeds into shorterwavelengths.

For nearly two decades, the Terahertz bands are beenefficiently used for imaging applications because these wavesare non-ionizing and able to penetrate through materials andabsorbed by water and organic substances. Their propertiesallow them to be used in communication networks to providehigher data rates up to Tbps. The Terahertz Gap is still the leastexplored band for its potential use in communication networksand to achieve higher data rates. Table IV, enlist the featuresof different frequency bands closest to Terahertz frequencybands. Its unique potentials motivate its usage for broadbandwireless communications.

B. Comparison between Terahertz band and other wirelesstechnologies

A brief comparison between existing wireless communica-tion technologies and Terahertz band communication includingmmWave communication is presented below and shown inTable IV.

1) Comparison with mmWave band communications :The mmWave bands comprises frequencies from 30 GHzto 300 GHz. Due to higher frequencies these bands facesevere attenuation due to oxygen absorption. Exceptionally,the propagation on 35 GHz, 94 GHz, 140 GHz, and 220 GHzexperience relatively small attenuation which enables long-distance communications between peers [53]. Other bands like60 GHz, 120 GHz, and 180 GHz attenuates up to 15dB/kmand also experience poor diffraction due to blockages [53].The range of mmWave and Terahertz can be reduced by pathloss, molecular absorption and atmospheric attenuation, theimpact on Terahertz signal is much more important than onmmWaves. The channeling effect can be compensated by usinghigh directional antenna for which mmWave is more maturethan Terahertz band. Additional features such as modulationand coding schemes, massive MIMO and phased antenna canenhance the spectral efficiency and transmission range. Anadvantage of Terahertz frequency over mmWave is that com-munication windows for Terahertz wave are higher than formmWave, and because of that Terahertz frequencies seem to bemore suitable for high data rate and low range communication.Progress on-device technology and physical layer for Terahertzsystems are still ongoing and for mmWave improved devicesand physical layer functionalities are realized.

The Terahertz and mmWave are neighboring bands buttheir properties are different. In comparison, the bandwidthat mmWave band is 10 GHz, which cannot support Tbpslink speed, whereas the Terahertz band has distance varyingtransmission windows of up to Terahertz bandwidth. To reachthe data rate up to 100 Gbps, the transmission schemes mustreach the challenging spectral efficiency of 14 bps/Hz [9].Moreover the link capacity required for few Gbps shouldbe several times higher than the user required data ratefor the timely delivery of data from multiple users. Withincreasing frequencies to Terahertz band, the Tbps links canattain moderate and realistic spectral efficiency of few bits persecond per hertz. The Terahertz band can also allow higherlink directionality compared to mmWave at same Transmitteraperture due to their less free space diffraction and shorterwavelength compared to mmWave. The transmitted powerand interference between the antennas can also be reducedby using smaller antennas with good directivity in Terahertzcommunications [54]. In Terahertz bands, the eavesdroppingchances are also lower compared to mmWave band due to highdirectionality of Terahertz beams, in which the unauthorizedusers must also be on the same narrow beam to interceptmessages.

The difference between mmWave and Terahertz bands aresummarised in Table V. From a technical point of view,Terahertz band offers much higher bandwidth than mmWaveband and therefore high data rate. The mmWave is deployed

4

TABLE II: General survey papers on Terahertz bands, devices, and communications.

Year Reference Network Area/Type Brief description of main topics covered2004 [36] Terahertz Communication Network An overview of communication and sensing applications is given with

sources, detectors, and modulators for practical Terahertz Communicationsystems.

2007 [30] Terahertz Communication Network The developments in the fields like Terahertz quantum cascade lasers,quantum well photodetectors, time-domain spectroscopy system and ma-terials are discussed with measurements of atmospheric propagation.

2010 [11] Nano Communication Network Propagation models for molecular and nano-electromagnetic communica-tions are discussed with challenges.

2011 [23] Terahertz Communication Network Different aspects of Terahertz communication are discussed with transis-tors, mixers, antennas, and detectors.

[29] Terahertz Communication Network Progress on Terahertz wave technologies is discussed.

2012

[33] Terahertz Communication Network Terahertz antenna technologies are discussed with different substrateintegrated antennas and beamforming networks.

[12] Nano Communication Network An overview on biochemical cryptography is discussed with requirementsrelated to security and challenges.

[37] Terahertz Communication Network An overview on demonstration of data transmission is given with stan-dardization activities.

[38] Terahertz Communication Network The Terahertz technology is discussed with challenges for spectroscopyand communications.

[16] Molecular communication networks The elementary models for intra body molecular communication channeland their extensions are discussed with challenges.

2013 [15] Nano Communication Network Issues of Nanonetworks are analyzed and discussed with particular focuson communication via microtubules and physical contact.

[39] Molecular Communication Network A review on bacterial communication and neuronal networks are givenwith application areas in body area networks.

2014 [9] Terahertz Communication Network Summarizes the research projects, spectrum regulation, and standardiza-tion effort for the Terahertz band.

2015 [19] Internet of Nano-Things A survey is presented for connecting body area networks and externalgateway for in-body nano communication. Network architecture, require-ments, and simulation based performance evaluation are also discussed.

[40] Terahertz Communication Network A survey on Terahertz technology is presented including devices, antennas,and standardization efforts.

2016 [32] Terahertz Communication Network A review is presented for impact of weather on Terahertz links, attenuation,and channel impairments caused by atmospheric gases like water vapor,dust, fog, and rain.

[31] Terahertz Communication Network A survey on graphene based devices for modulation, detection, andgeneration of Terahertz waves is discussed.

2018[24] Terahertz Communication Network A review is presented for channel modelling for Terahertz band including

single antenna and ultra massive MIMO systems.[41] 5G Femtocell Internet of Things A survey on low Terahertz band circuit blocks is presented with focus

on energy consumption using best modulation schemes and optimizinghardware parameters.

[5] Terahertz Communication Network A review is presented for deployments of Terahertz wireless link andopportunities to meet future communication requirements.

2019 [42] Nano Communication Network A summary of current status of Nano communication network is presentedwith applications and different layers of protocol stack.

2019 [43] Terahertz Communication Network Discusses wireless communications and applications above 100 GHzbands with opportunities and challenges.

2019 [44] Terahertz Communication Network Recent activities on Terahertz development, standardization, applicationsand communications are reported.

2019 [45] Terahertz Communication Network Discusses wireless communications and applications above 100 GHzbands with opportunities and challenges.

2019 [46] Terahertz Communication Network A survey on Terahertz communications, applications and layers of protocolstack is presented.

2019 [47] Terahertz Communication Network A survey on Terahertz communications, applications and layers of protocolstack is presented.

2019 [48] Terahertz Communication Network A review is presented on development towards Terahertz communicationswith key technologies.

5

TABLE III: Survey papers discussing the Terahertz MAC layer.

MAC protocolsclassification

MAC functionalities

Network type Year Reference MA

Cla

yer

disc

usse

d

App

licat

ion

area

s

Cha

lleng

es

Net

wor

kto

polo

gies

and

scal

e

Rx/

Tx

Initi

atie

dco

mm

unic

atio

n

Cha

nnel

acce

ss/s

hari

ng

Inte

rfer

ence

Err

orco

ntro

l

Pack

etsi

ze

Dev

ice

disc

over

y

Han

dsha

king

Tx

dist

ance

Dat

ara

te

Ant

enna

s

Bea

mfo

rmin

g

Mod

ulat

ion

Cro

ssla

yer

Nanonetworks

2008 [49] Partially X X X X X X X X X X X X X X X X2010 [14] Partially X X X X X X X X X X X X X X X X2010 [17] Partially X X X X X X X X X X X X X X X X2010 [50] Partially X X X X X X X X X X X X X X X X2011 [10] Partially X X X X X X X X X X X X X X X X2012 [51] Partially X X X X X X X X X X X X X X X X2012 [13] Partially X X X X X X X X X X X X X X X X2016 [18] Partially X X X X X X X X X X X X X X X X2017 [52] Partially X X X X X X X X X X X X X X X X

Vehicular networks 2017 [21] Partially X X X X X X X X X X X X X X X X

Terahertz networks

2014 [3] Partially X X X X X X X X X X X X X X X X2014 [4] Partially X X X X X X X X X X X X X X X X2016 [28] Partially X X X X X X X X X X X X X X X X2016 [27] Partially X X X X X X X X X X X X X X X X2019 [35] Partially X X X X X X X X X X X X X X X X

This work 2019 Detailed X X X X X X X X X X X X X X X X

actually in WLAN as well as cellular network system bycontrast to THz where researchers are striving to designnew devices with high power generation and low receivingsensitivity. Due to the larger coverage and mobility support,mmWave communications can be used in backhaul commu-nication and cellular communications. Whereas Terahertz canbe used where high throughput and low latency are requiredwith fixed infrastructure for now until mature devices willbe available. MmWave antenna is much more mature thanfor THz, it is possible to deploy antenna diversity and beamsteering and tracking for mmWave. The channel model formmWave is well developed as measurements are carried formany mmWave windows as well as for different scenarios. ForTerahertz band few measurement campaigns are performedparticularly for indoor scenarios around 300 GHz and 100GHz. The free space attenuation increases as a function offrequency and molecular absorption loss occur due to oxygenmolecules in mmWave, whereas in Terahertz band it occursdue to water vapors. The reflection loss is high for bothmmWave and Terahertz band which results in severe loss ofNLOS path compared to LOS path. The scattering effect alsobecomes severe when the wavelength decreases below 1 mmwhich results in increase of multipath components, angularspreads, and delay. Due to much smaller wavelength many an-tennas can be packaged together to generate narrower beams.However, the stronger directivity increases the difficulties andoverhead of beam alignment and tracking but reduces theinterference. The difference between mmWave and Terahertzband communication can also be found in [35], [55] for furtherreading.

2) Other technologies: The traditional 802.11 protocol ismainly designed for 2.4 GHz WiFi, which uses frequency-hopping spread spectrum and direct sequence spread spectrum.It provides a simple data rate of up to 2 Mbps. After that802.11 (a and b) were published, operating at 5 and 2.4GHz bands. The 802.11a is based on Orthogonal FrequencyDivision Multiplexing (OFDM) and can provide a data rate

of up to 54 Mbps, whereas 802.11b supports only 11 Mbps.The 802.11 ac was published which aimed at providing thedata rate up to more than 100 Mbps. Other than those, the802.11ad is developed for the carrier frequency of 60 GHzand it belongs to mmWave frequency bands. The detaileddescription and comparison are provided in Table IV. Thesmart technologies like OFDM and communication schemeslike large-scale MIMO can be used for frequencies below5 GHz to achieve higher spectral efficiency. In Long-TermEvaluation Advanced (LTE-A), the peak data of up to 1 Gbpsis possible only when a 4x4 MIMO scheme is used over a 100MHz aggregated bandwidth [56].

The frequency bands above 10 Terahertz cannot supportTbps links. Although very large bandwidth is available in FSOcommunication system which operates at IR frequencies, itstill holds some issues which limit its use for personal wirelesscommunication like the atmospheric effects on the signalpropagation (fog, rain, pollution and dust); high reflection loss;misalignment between transmitter and receiver; and low powerlink budget due to health safety which limits both transmissionrange and achievable data rates for FSO communication. Itcan support up to 10 Gbps of data rate with a proper line ofsight (LOS) for Wireless Local Area Network (WLAN) [57].For non-LOS much lower data rate has been reported [58].For longer distance, an FSO system was demonstrated in [59]to support 1.28 Tbps, however, requires typical fiber op-tical solution to generate and detect high capacity opticalsignals which are injected in the optical front-end and alsodoes not include the signal generation, detection, modulation,and demodulation blocks. These constraints limit the overallfeasibility to achieve higher data rates for 5G and beyondnetworks. A comparison of wireless and optical technologiesis presented in [2] for wireless indoor environments. It ismentioned in [2], wireless communication has overall betterchances for penetration through obstacles in comparison withFSOs. Further, the IR and ultraviolet are not considered as safefor human and the Visible Light communication requires the

6

visibility of light at all times. A comparison of visible lightand sub-Terahertz band communication is discussed in [60]for 5G and 6G wireless systems.

C. Background and motivation for Terahertz MAC protocolsIn this Section, the definition of MAC layer protocols is

defined first with motivation for the need of different MACprotocols for Terahertz communication networks is presented.

1) Medium Access Control background: The MAC layeris mainly responsible for controlling the hardware to interactwith wireless transmission medium by flow control and mul-tiplexing. It serves the interaction between the Physical andUpper layers. It provides a method to identify the frame errorsand sends the packet to Physical layer when channel access ispermitted or scheduled. It also decides when and how muchto wait before sending a packet to avoid collisions among thenodes. Different wireless technologies require different MACprotocols to serve the transmission purpose. For example, theMAC protocols for LTE and GSM standard has different userrequirements to serve than the required MAC functionalitiesby a Wireless sensor network.

As the user demands and network requirement are enhanc-ing, efficient MAC protocols are in demand to assist thenetwork operations and provide adaptive solutions. The MACalthough suppose to provide same functionalities mentionedabove, but due to different band features and requirements(cf. Section IV), new mechanisms are required to facilitateuser demands and networks requirements (cf. Section III).

2) Review of various Terahertz MAC protocols: Designingof an efficient MAC protocol for Terahertz wireless communi-cation is crucial for future high-speed networks. The existingwork on Terahertz MAC protocols can be broadly classifiedbased on network topologies as centralized, clustered and dis-tributed (Adhoc network); Channel access schemes as random(contention-based), scheduled (contention free) and hybrid;handshake mechanism as receiver and transmitter initiatedcommunication; and applications as macro and nanoscale, asshown in Figure 2.

The network topologies have a great interest which canaffect the MAC protocol design and they are discussed in Sec-tion V. In centralized networks, a controller or an access pointis used to provide coordination among the nodes, whereasin distributed networks each node takes its own decision andcoordinates in a distributed manner. In cluster-based networks,nodes form a group to transmit their information via a singlegroup head.

The channel access mechanisms are presented in Section VIand they include mainly random, scheduled and hybrid channelaccess mechanisms. In random channel access, each nodecontends for a shared channel to improve channel utilizationand can introduce collisions. The directional antenna usagein Terahertz frequency band in carrier sense based multipleaccess can provide spatial reuse of spectrum which can im-prove the throughput of the network. Whereas, the schedulebased channel access methods like TDMA assigns channelaccess slots in advance without spatial reuse. This providesa contention-free environment at the cost of high synchro-nization overhead. In [67], the scheduled based mechanism

is shown to provide worse throughput and latency perfor-mance compared to random schemes. The hybrid channelaccess mechanism combines the functionalities of random andscheduled channel access mechanisms. In this scheme, thechannel access is performed using random scheme followedby scheduled data transmission.

The initial access mechanisms are presented in Section VII.The unique features of the Terahertz band (discussed inSection IV-A) adds different challenges for Terahertz wirelesscommunications like directional antenna and high path loss.These unique features, network scale, and application require-ments demand new initial access mechanisms. The receiverinitiated communication is mainly followed in networks withsevere energy limitation, in which a receiver triggers the com-munication when it has sufficient energy to receive a packet.Whereas, transmitter initiated communication is mostly fol-lowed in wireless communication, in which a sender whenhaving some data to transmit triggers the communication.

Moreover, the Terahertz MAC protocols can also be clas-sified based on applications as nano and macro scale ap-plications. These applications are discussed and their designchallenges are identified in Section III. The nano applicationsinvolve applications with range up to a few centimeters,whereas applications with range higher than a meter can beconsidered as macro scale applications.

3) Difference between Terahertz MAC protocols and proto-cols for other communications: Terahertz is characterized byhigh bandwidth comparing to lower frequency band, it is alsocharacterized by its high-frequency attenuation. Therefore,traditional concepts of MAC should be modified or extendedin most cases to answers application requirements such as ultrahigh throughput, coverage and low latency. The Terahertz com-munication allows concurrent narrow beam links as comparedto the traditional networks which is an interesting researchtopic for Terahertz communications including link schedulingfor interference mitigation and capacity management in MAClayer. The unique features of the Terahertz band, and TerahertzMAC protocol design considerations are discussed in detail inSection IV. The differences with traditional wireless MAC arepresented below.

Link preparation for access, ACK and Data: The traditionalWiFi networks use in-band data and ACKs with short inter-frame spacing gap of 20 ms between data and ACK message.In Terahertz using the same link for ACK and data involvesmany challenges including, beam switching and steering, andadditional synchronization overhead, which can increase thedelay.

Deafness issue: The traditional networks operates in bothomni and directional antenna modes, whereas in Terahertznetworks directional antennas are required at both the senderand receiver, which can introduce deafness and collisionproblem. Because the nodes can sense only in one direction ata time which makes it difficult to capture accurate and timelyinformation of their neighbors.

Noise and interference : Terahertz band is sensitive tothermal and molecular noise, interferences come generallyfrom receiver side lobe and interferers beam alignment orfrom multipath reflections. It can be neglected if directional

7

TABLE IV: Different types and features of wireless communication technologies [5].

Technology Frequencyrange

Wavelength Data rate Transmissionrange

Power con-sumption

Topology LOS/nLOS Noisesource

Weathereffect

mmWave 30GHz-300GHz

3cm-1mm 100 Gbps [7]–[9] High Medium PTP,PTmP

both Thermal Sensitive

Terahertz

100-300GHz (sub-THz)300GHz-10THz

1mm-30µm

upto 240 GHZ: 10 Gbps [61]upto 300 GHz: 64 Gbps [62]300-500 GHz: 160 Gbps (sin-gle channel) [63]300-500GHz: > 160 Gbps(multiple channels) [64]

Short/Medium (1-10m) Medium PTP, PTmP both

Thermalandmolecularnoise

Sensitive

Infrared 10THz-430THz

30µm-3µm

2.4 kbps to 1 Gbps Short, upto 1 m Low PTP LOS Sun/Ambient Sensitive

VisibleLight

430THz-790THz

0.3µm 100 Mbps to Gbps [65],[66]

Short Low PTP both Sun/Ambient

Ultra Vio-let

790THz-30PHz

100-400 nm Short Low PTmP Sun/Ambient Sensitive

Terahertz MAC protocols classification

Network Topologies(Sect.V)

Centralized(Sect.V-A)

Clustered(Sect.V-B)

Distributed(Sect.V-C)

Channel Access Mechanisms(Sect.VI)

RandomChannel Access

(Sect.VI-A1,VI-B1)

SceduledChannel Access

(Sect.VI-A2,VI-B2)

HybridChannel Access

(Sect.VI-B3)

Tx/Rx Initiated Communication(Sect.VII)

Tx Initiated(Sect.VII-A)

Rx Initiated(Sect.VII-B)

Fig. 2: Classification of Terahertz MAC protocols based on network topologies, channel access mechanisms and Tx/Rx initiatedcommunication. They are further discussed based on Terhaertz applications and network scale as macro and nano.

antennas are used, but still can reduce system performances,mmWave is much more mature regarding interference andnoise modeling, techniques to reduce impact on the systemare available using adequate equalizers and filters as well ashigh-efficiency coding schemes. For Terahertz system workis ongoing on techniques to mitigate interference and reducenoise effect. From MAC point of view designing efficientscheduling algorithm can improve system performances.

Antenna directionality and beam management: In tradi-tional networks mostly Omni-directional antennas are used.Although directional antennas are used for point-to-pointconnectivity, they cover high transmission distance. However,using directional antennas for Terahertz band, introduce manychallenges including beam management for initial access anddata transmission, frequently than in traditional networks. Fastbeam tracking mechanisms are required mainly at Terahertzband involving applications with mobility.

Frame length and duration: Long frame can increasethroughput however frame error rate can increase, and anefficient error control mechanism should be used to mitigateframe losses. For Terahertz system, the frame duration is verylow compared to microwave and mmWave system. The systemcan accept more users without affecting overall delay, forexample the scheduling time for LTE at microwave frequencyis equal to 1 ms whereas for Terahertz it can be much lower,and more nodes can join system within 1 ms, if we consider itas a threshold for network delay. In order to increase nodes in

the system, fast beam switching and steering is then required.For frame size, as Terahertz system is still immature, framesshould be of low size, as they are prone to channel errors.Therefore robust channel coding should be implemented todetect and correct bits and reduce retransmissions and delays.Comparing to traditional techniques, it is possible to re-usetraditional mechanisms for frame error handling such as frameretransmission.

Channel access and huge bandwidth availability: Thereis a huge bandwidth availability in Terahertz bands up toseveral GHz, which needs to be managed to support veryhigh throughput and ultra-low latency at MAC layer. However,the bandwidth available for tradition networks, as discussedin previous subsection, is few MHz or up to a few GHz incase of mmWave band communication. The huge bandwidthavailability means less contention and collision at these bandswith low transmission times. The main functionalities requiredat Terahertz band are coordination and scheduling, whereas intraditional networks contention and interference managementare main requirements for channel access. Therefore, TDMAwith highly directional antennas can be a good choice forTerahertz networks.

High energy fluctuations: The heavy energy fluctuations atnanoscale networks require efficient energy harvesting mech-anisms in support of MAC layer to perform handshaking anddata transmissions. Whereas in traditional networks energyfluctuations are not as high as in nanonetworks.

8

Scheduling: For macro network, MAC is based on timedivision techniques in most case as communication can occuronly if beams are aligned between nodes, the scheduling isa crucial part of the THz MAC procedure that makes THzmacro networks different from traditional MAC techniques.

Throughput and latency requirements: Traditional networksoffer a limited amount of throughput, as discussed in previoussubsection. Higher throughput is possible by using MIMOand higher-order modulation techniques, whereas, in Terahertznetworks, low complexity modulation techniques are requiredfor nanoscale networks. The latency requirements are also verytight in Terahertz networks.

III. TERAHERTZ BAND APPLICATIONS AND THEIRREQUIREMENTS

The Terahertz band has the potential to support futureapplications with high throughput, low latency, and massiveconnectivity. Therefore, it is vital to explore the possibleapplications and extension of existing applications to deliverhigh-quality transmission including voice, video, and data,over Terahertz band communications. The existing applica-tions are categorized here as macro and nano applications.Typically, these applications include outdoor as well as in-door applications that require speed from Gbps to Tbps.There are applications in which users require Gbps speedlike small cells, WLAN, vehicle to vehicle communicationand device to device communication. The applications whichrequire Tbps speed and can not be satisfied using traditionalbands include applications that use traffic aggregation andnano communication, for example backhaul communication,edge communication within a Data Centre and nanodevicescommunication which can utilize full bandwidth of Terahertzband due to small distance. The Terahertz link can be usedto aggregate 5G traffic including control plane signaling, IoTtraffic, internet, and mobile services at the backhaul and corenetwork side to replace existing optical fiber links. It can beused also for traffic augmentation. The Tbps links can alsobe required for inter-chip communication, where chips canexchange ultra-high data rate in a flexible way using short-range, then THz communication will be feasible with the ultra-high data rate. These applications are also shown in Figures 3and 4. Their design requirements are highlighted to emphasizetheir particular necessities to progress in the Terahertz MACprotocol design. Their performance target requirements aregiven in Table VI. Table VII, presents the details of theseprotocols based on different communication aspects and pa-rameter aware Terahertz MAC protocols1. The MAC layerrelated design requirements, issues and considerations will bediscussed in Section IV.

A. Applications for Macro Scale Terahertz NetworksThe macro-scale communication involve applications in

which the transmission range is higher than 1 meter and

11) Channel aware: Nodes are aware of the spectrum information, 2)Physical layer: Nodes are aware of physical layer parameters like propagationloss and bit error rate, 3) Memory aware: Nodes are aware of the availablememory at each node, 4) Position aware: Nodes are aware of the position ofother nodes, 5) Nodes are aware of the bandwidth and adapt according to theavailable bandwidth.

up to a kilometer. The Terahertz bands although has hugebandwidth availability, the transmission distance can vary dueto high path and absorption loss. The indoor applicationdiffers from outdoor applications mainly due to scatteringand reflection effects. Therefore requires different channelmodels for different environment which should be consideredwhile designing MAC protocols for these applications. Theseapplications are shown in Figure 3. Table VI, is showing thetechnical requirements for different Terahertz applications. Itincludes fixed point-to-point, point-to-multipoint, and mobilityscenarios.

The applications which require Tbps links include wirelessbackhaul [111] and Data Centres [112]. Wireless Backhaultypically involves point-to-point connections for informationtransmission to the base stations of the macrocell, especiallywhere the fiber optic is not available. Small cell communica-tion can be another application scenario in which the backhaulpart will require Tbps links to transfer huge amount of data.High power antennas can be used for larger distance coveragebut should consider environmental factors and frequent userassociation while designing a MAC protocol. The atmosphereloss is higher, however can be compensated by high antennadirectivity. For capacity enhancement and larger bandwidthwith Tbps transmission, the Terahertz band between 200and 300 GHz has shown low atmospheric losses [113]. Thewireless fiber extender is also an interesting scenario to extendthe communication range and capacity of existing backhaulcommunication set up to provide reliable data communicationwith Tbps throughput for distance up to 1 Km in outdoorenvironments. A very large antenna array or Massive MIMOtechniques can be used to transfer information between cells.The use of massive MIMO arrays can be used in an adaptivemanner to modify the transmit and receive beams to accom-modate the changes in the environment than to physicallyreadjusting the arrays [100]. They can be used to communicatewith multiple backhaul stations by electronic beam steering.The challenges for MAC include reliable data transfer, interfer-ence, and atmospheric loss mitigation techniques and adaptiveparameter adjustment including frequency switching, distance,and bandwidth allocation mechanisms.

At the macro scale, another promising application is wire-less Data Centres [2], [114]. In an attempt to increase the userexperience and high-speed network, the cloud applicationshave introduced the competition between the Data Centres.This resulted in extensive extension of servers and requiredbandwidth to support numerous applications to manage burstyhigh traffic. The Terahertz band can be augmented to supportTbps links especially at the edge where the traffic aggregatesrather than using the cables with limited capability. The Ter-ahertz links can be used in parallel with existing architectureto provide backup links, failover mechanisms and high-speededge router links with SDN support [115]. This can improvethe user experience and can also reduce the cost of deploymentwithin Data Centres.

Using the high capacity wireless Terahertz links can alsohelp in re-designing the Data Centres geometry [116]. How-ever, require careful communication protocol design like Phys-ical, MAC and network layer, to be efficiently utilized for

9

TABLE V: Comparison between mmWave and Terahertz wave technologies.

Milimeter wave technology Terahertz wave technologyTransceivers Device High performance mmWave devices Avail-

able [68]Available immature: UTC-PD, RTD andSBD [3]

Modulation and coding High order modulation techniques avail-able [69]. For example, QPSK, LDPC 1/2(264 Mbps) 16QAM, LDPC 1/2 (528 Mbps)

Low order modulation techniques (OOK,QPSK) [3], LDPC, Reed Soloman, Hamming

Antenna Omni and high gain directional, MIMO sup-ported [7], antenna gain =18dBi when 8X8antenna array used [69]

Omni and Directional, phased array with lownumber of antenna elements (4x1) [70]

Bandwidth 7GHz @60GHz [43] 69 GHz at 300GHzchannel models [71] Yes PartiallyData rate Maximum of 100Gbps [7], [8] 100 Gbps [44], [72] to few Tbps (theoretical)Standards 5G NR, IEEE 802.11 ad, IEEE 802.11 ay IEEE 802.15.3dMobility Supported [73] Not yetBeam management [74] Yes NoAdaptive beam searching andswitching time

45 ms [69] in progress

Outdoor deployment Yes NoFree space loss Low HighCoverage High [75] LowRadio Measurements [71] Available for many windows: 28GHz, 72GHz,

52 GHz, 60 GHz300 GHz indoor, example measurement car-ried at data center environment

Device size Few millimetres Few micrometersEnd to end simulators Available on ns3 for 5G cellular network [76],

[77]NS3 Terasim [78]

Fig. 3: Terahertz communication applications for macro scale networks.

the Data Centre network. The top-of-rack (ToR) Terahertzdevices can connect using point-to-point and multipoint wire-less links. However, requires directional antennas for interrack communication for enhanced coverage for more thana meter. The intra rack communications can also use omni-directional antennas due to short distance between the routers.A fair and efficient channel access scheme is required for bothinter/intra rack communication with scheduling (for directionalantennas) and with collision avoidance (CA) techniques (forOmni-directional antennas) due to multi-user interference.To connect different ToR devices, the link establishment isvery important however becomes challenging with directionalantennas and energy minimization constraint.

The other point-to-point connectivity applications includethe KIOSK downloading system which can be used for instanttransfer of bulk amount of data, as shown in Figure 3. Its a peerto peer communication system between stationary transmitterand a mobile receiver with limited mobility. It can be imaginedas a stationary terminal or point, which may be connectedwith a data center using fiber-optic and can provide the bulkamount of data to various users in few seconds, typically inGBs. This type of application can use Tbps links to satisfyuser demands. The challenges include rapid user associationand link establishment, beam management, error detection,and correction strategies, and secure authentication on publicsystems. An experimental demonstration of a prototype for

10

TABLE VI: Terahertz applications features, requirements and some general MAC related challenges.

Networkscale

Category Application Ar-eas

Coverage Mobility Datarate

Latency No of con-nections

Linkavailabilityand reliability

Connectivity Energyefficiency

TargetBER

General and MAC related challenges

Macro

Indoor applications

Data Centre net-works [2], [22],[79]

< 20 m No upto100Gbps

0.1-0.2ms

large 99.99% P2P,P2MP Low pri-ority

10−12 Efficient and shared channel access, error andflow control with link reliability, coverage,device discovery, beam management, cover-age, high throughput and low latency, resourcebased link assignment, channel modeling toconsider interference management for MAC.

TerahertzLocal AreaNetworks [22],[25], [80], [81]

< 50 m Yes upto100Gbps

< 1 ms medium High P2P, P2MP,Adhoc

Mediumpriority

10−6 Efficient channel access, user mobility anddensity support, interference and collisionavoidance, high reliability and throughput withthe low latency, coverage, user association andbeam handovers.

KIOSKdownloadingsystem [22],[81]–[83]

0.1m No 1 Tbps < 1 ms small 99.99% P2P Low pri-ority

10−6 Error correction algorithms, fast dadta deliv-ery, user association and disassociation, shortterm link establishment.

TerahertzPersonal AreaNetworks [22],[25], [80], [84]

< 20 m Yes upto100Gbps

< 1 ms meidum High P2P, P2MP,Adhoc

Mediumpriority

10−6 Scalability, interoperability between multipledevices, high throughput and very low latency,mobility and coverage.

InformationBroadcast [85],[86]

0.1-5m Yes 1 Tbps upto fewsec

small tomedium

99.99% P2P,P2MP Low pri-ority

High throughput and low latency, directionaltransmission, synchronization, error correc-tion, mobility and coverage.

Outdoor applications

Small and ultradense cell tech-nology [20], [22],[87]–[90]

10-15m

Yes >100Gbps

upto fewms

medium tolarge

High P2MP High pri-ority

10−10 High user mobility, frequent handovers, seam-less connectivity, high data rate support, verylow latency, ultra high reliability, usage of di-rectional antennas, beam alignment and man-agement.

Vehicularnetworks anddriver-lesscars [21], [90]–[94]

> 100m

Yes >100Gbps

upto fewms

medium medium P2P, P2MP,Adhoc

High pri-ority

Data cache system and scheduling, realtime data transmission and optimization, au-tonomous and timely data delivery.

Military applica-tions [95]–[97]

> 100m

Yes 10-100Gbps

upto fewms

medium High P2P, P2MP,Adhoc

For fixedunits:Lowpriority.Forsensors:Highpriority

Security for data, user and to avoid jammingattacks, last mile connectivity, larger area cov-erage.

Space Appli-cations [98],[99]

kms Yes 10-100Gbps

upto fewsec

small tomedium

High P2P,P2MP Low pri-ority

Modulation and multiplexing techniques, freespace communication, signal attenuation studyfrom space to ground communication, efficientfrequency utilization, new antenna design.

Backhaulconnectivity[22], [25], [81],[100]

1 km No >100Gbps

upto fewms

small High P2P Low pri-ority

10−12 Distance dependant bandwidth adjustment,beam management, link reliability, adaptiveparameter adjustment.

Nanoscale

In/On-bodyapplications

Health monitor-ing [49], [101],[102]

0.5-2.5mm

Yes 100Gbps 1 ms large High adhoc,P2MP

High pri-ority

Efficient usage of Terahertz frequencies, safetyconstraints and heating problems, interac-tion between nano devices and surround-ing environment, hybrid nano communica-tion system, efficient communication proto-cols, nanonetwork architecture, antenna de-sign, channel/propagation model.

Outdoor applications

Defence applica-tions [103]

No 100Gbps 1 ms large High adhoc,P2MP

High pri-ority

Timely dissemination and gathering of data,Multiple access due to molecular communica-tions, communication range, device scalability,channel/propagation model.

Agricultural ap-plications [104]–[106]

4mm No 100Gbps 1 ms large High adhoc,P2MP

High pri-ority

High resolution monitoring of chemical, com-pound emissions from plants, obstacles detec-tion and avoidance, efficient communicationprotocols.

Indoor applications

Internet of Nano-Things [19],[107]

2m No 90Gbps 1 ms massive High adhoc,P2MP

High pri-ority

Service or device discovery, information rout-ing, reliability, channel sharing.

Intra chipnetwork [22],[81], [100],[108]–[110]

0.03m No 100Gbps 1 ms small tomedium

High Ad-hoc Mediumpriority

10−12 Inter-core communication, high bandwidth,low delay.

KIOSK Downloading system using 300 GHz band is presentedin [83] and [117], which includes the channel and LOSanalysis with comparison of error correction algorithms. Theapproximate beam size is mentioned in [83] as 22 cm with 30dBi as antenna gain for 1-meter distance.

Vehicle to infrastructure and its backhaul network canalso utilize Tbps links to improve vehicular communicationnetworks. For example, Google’s auto-driver cars generate thesensor data at the rate of 750 MBps [118] and are expectedto generate 1 TB of sensor data in a single trip [119]. The

delivery, processing, and optimization of such huge data toassist vehicles can require Tbps links to improve the efficiencyand latency of communication. As an alternative solutionfor fiber-based backhauling, the vehicles can also serve asdigital mules to reduce the deployment cost and to migratethe data [120], [121].

The point-to-point communication can also be utilized inmilitary applications between air to air and ground ma-chines [96], [97]. For many years, space agencies such asNASA and ESA, are developing sensors and instruments for

11

Fig. 4: Terahertz communication applications for nanoscalenetworks.

space technology [98], space applications range for point-to-point communication [99], [122], [123]. Using THz link forintra satellite application can be feasible outside the atmo-spheric region, where the only free space loss is considered. Apossible scenario can be an inter-satellite link within the sameconstellation or between low orbit and geostationary satellite.

The interesting scenarios for TLAN are the indoor homenetworks and LAN [80], [124]. Mainly, the backhaul part ofthis home network to core networks can use Tbps links totransfer aggregated data, which can facilitate multiple users ata time to download huge data. However, currently, the shortdistance within an indoor home or office environment requirespoint-to-point communication or efficient beam managementstrategies. The high speed and instant connectivity betweenmultiple personal devices are possible using Terahertz com-munication links [80].

1) Summary of macro scale applications: The Terahertzapplications range from indoor to outdoor from short to largedistances. The applications with very short communicationdistance like KIOSK downloading system and intra rackcommunication within a Data Centre does not involve mobilitychallenge. However, a delay can be involved depending uponthe amount of data transfer and channel access schedule.Mainly, they require point-to-point connectivity between thesystems. The last mile access and backhaul point-to-point andmultipoint are also very interesting scenarios that require highthroughput and low latency for longer distances up to a km.Currently, it has reached 10 Gbps and is realized to reachbeyond 100 Gbps using Terahertz band with high bandwidthavailability [125], [126].

The atmospheric losses affect both indoor and outdoor typeof applications differently and therefore the appropriate chan-nel and propagation models should be considered. The indoorenvironments like Data Centres require fixed links between theracks, very limited mobility. The outdoor environments likebackhaul links involve fixed point-to-point links, however dif-fer from indoor environments in terms of atmosphere, distance,reliability and link-budget requirements. The scenarios likevehicular and small cells require high mobility and thereforeinvolve frequent handovers and must support high user density

and scalability for new MAC protocols. For both scenarios,efficient channel access mechanisms, reliable connections witherror detection and correction are required with efficient beammanagement techniques.

B. Applications for NanoScale Terahertz Networks

Nanotechnology enables the nano-components which areable to perform simple specific tasks, like data storage, com-putation, sensing, and actuation. These nano-components canbe integrated into a device with only a few cubic metersin size and can enable the development of more advancednano devices [3]. These nanodevices can exchange informationto achieve complex tasks in a centralized or a distributedmanner [127], which enable unique and interesting applica-tions in the field of biomedical, industrial, military, healthmonitoring, plant monitoring, nanosensor networks, chemicaland biological attack prevention and system-on-chip wirelessnetworks [3], [101], [128]. It mainly involves communicationupto few centimeters. This also includes the networks usingelectromagnetic radiation like nanonetworks and molecularnetworks in which transmission occur using the flow ofmolecules [101], [128].

The nano devices, due to their smaller size and communica-tion range can benefit with larger bandwidth and so can utilizeTbps links. Due to smaller range, the path loss remains at thelowest which enables high throughput in nano communication.These applications are shown in Figure 4 and their challengesare also mentioned in Table VI. The nanosensors can beused to detect an early disease by using molecular commu-nication by gathering heart rate. The gathered informationthat can be transmitted over the Internet using a device ormobile phone to a healthcare provider. Other applicationsare nano-scallop to swim in a biomedical fluid [129], a bio-bot which is powered by skeletal muscles [130], on boarddrug delivery system [131], a magnetic helical micro-swimmerto deliver single cell gene to the human embryonic kidneyin a wireless way [132]. The MAC layer protocols shouldconsider energy consumption and harvesting trade-off, errorrecovery, scheduling of transmissions among different nodesand efficient channel access.

Terahertz band can be used for wireless on-chip net-works [109], [133]. It can be used at the very small scale toconnect two chips due to its high bandwidth and low area over-head. The MAC should consider supporting maximum num-ber of cores by addressing MAC performance by specifyinginput traffic and interface characteristics [108]. The tolerabledelay should be analysed among different layer architectureand analysis of maximum cores supported for throughputdelay [108]. The challenges include efficient channel accessmechanism for intra chip communication with scheduling,efficient inter-core communication, small-scale antennas toprovide high bandwidth and low delay.

1) Summary of Nanoscale applications: The nano appli-cations, especially on chip communications can utilize highbandwidth and so can provide Tbps links. In nano com-munications the communication range is very small, due towhich the path loss remains at the lowest. Mainly, these

12

applications requires efficient energy consumption and har-vesting mechanisms to address limited energy issues whileconsidering nonbatteries/nanogenerators/nontransreceivers ar-chitecture and performance enhancements. The timely dis-semination of data from nano sensors to external network.The antenna technology and new channel/propagation andnoise models are required with tools to estimate path loss fordifferent nanoscale network environments. Efficient commu-nication protocols such as modulation and coding techniques,power control, routing and MAC strategies for nanoscale com-munications. Their MAC protocols also requires to supportscalability and connectivity among large number of devices.Due to limited capacity of devices, energy efficient mechanismare required with harvesting and transmission balance forefficient communication. Link establishment and schedulingthe communication among devices is also a challenge. Somearchitectures for nanonetworks to handle complex task bycombining it with current technologies like SDN, IoTs, virtualnetwork and fog computing are presented in [18].

C. Other applications for Terahertz communications

The Terahertz band is emerging as the most promising wayto realize Gbps link, due to increase in the current trafficdemand [100]. A radio link over 120 GHz band for datatransmission at the rate of 10 Gbps to the range of over 800m is presented in [134]. There are applications mostly at theend users, which requires high speed data connections upto amore than 100 Gbps [27], [85]. That can be fulfilled by usinglow Terahertz frequency bands with higher order modulationschemes. These applications include a broadcast system atpublic places like metro stations, airports, and shopping mallsto transfer data in GBs. However, the transmission or datatransmission currently can be possible only for the shortdistance. The short distance, mobility and user density percoverage area should be considered while designing a MACprotocol. The benefits of this application is also recognised byIEEE 802.15.3d (TG - 100 Gbps wireless) as one of the usecases for the Terahertz communication [86]. An informationbroadcast case with area density, number of users and mobilitypattern is discussed in [85]. Further, the front ends of applica-tions like Terahertz LAN, vehicle to vehicle communication,small cells to mobile users are all applications where Terahertzbands can be used for high throughput. However will requireextreme antenna directionality management and rapid userassociation and disassociation mechanisms.

The fronthaul is between the base station and the enduser radio equipments which requires Gbps links. These radioequipments can be mobile as in small cells and wireless LANscenarios and can be fixed users. The critical parameters forthese applications other than high data rate (upto 1 Tbps) isdistance which should be above a meter to a kilometre. Futureapplications which will include massive deployment of smallcells for cloud radio access network which may increase thedata rates for end users at front end and back end which willrequire Tbps links. The small cells deployment can utilize thehuge bandwidth available in Terahertz band and can free upthe lower frequency bands which leads to several Tbps of data

transfer [27]. One of the possible and upcoming applicationsof Terahertz band is the small cell communication for mobilecellular networks, in which ultra-high data rate can be providedto mobile users within transmission range up to 20 m [56],[90]. The Terahertz small cell can be a fixed point installedto serve multiple mobile users. The mobility of users withhigher data volume offloading needs to be supported. Theusers moving from cell to cell requires seamless handoverfor uninterrupted communication. The Terahertz directionalantenna usage can increase challenges in user associationand tracking with scheduled channel access. Therefore, theTerahertz MAC protocol should consider these requirementsand the target performance to ensure the user satisfaction.

The virtual reality (VR) device is an interesting applicationwhich requires at least 10 Gbps data traffic transfer. However,currently it relies on wired cord and needs to be shiftedto wireless transfer with more than 10 Gbps data rate. TheVR applications mainly requires ultra high reliability andlow latency with high data rates for their services and fastdata processing [135]. Using Terahertz band for VR servicesrequires transmission and processing delay to be low.

IV. DESIGN ISSUES AND CONSIDERATIONS FORTERAHERTZ MAC PROTOCOLS

This section discusses the feature of the Terahertz bandwhich needs to be considered while designing efficient MACprotocols, the design issues and challenges related to TerahertzMAC protocols.

A. Feature of Terahertz band communication related to Tera-hertz MAC protocol design

By using frequencies above 0.1 THz, new propagationphenomena can appear such as reflection, wave absorptionby some molecules and channel noise generation [169]. Theunderstanding of the Terahertz band seems to be crucial todesign systems exploiting this frequency, hence, researchersare focusing on the behavior of Terahertz wave travelingunder different environments and in the presence of itemssuch as walls, concrete or grass. Following are the Terahertzband features which can affect the MAC-layer performanceincluding throughput and delay.

1) Path loss: To realize the Terahertz band and its char-acteristics, it is important to understand its propagation phe-nomenon and to analyze the impact of molecular absorptionon the path loss and noise [170]–[172]. The current effortsare mainly focused on channel characterization at 300 GHzband [173]–[177]. As Terahertz wave propagates, it suffersfrom different types of attenuation due to absorption, scat-tering, and refraction [178]. It can follow different paths atthe receiver as the sum of non-/line of sight. The path lossincludes the spreading as well as the absorption loss. Thespreading loss occurs due to the expansion of waves as itpropagates through the medium, whereas the absorption lossoccurs when a Terahertz wave suffers from the molecularabsorption at the medium [179]. These losses make a Terahertz

13

TABLE VII: Summary of existing Terahertz MAC protocols with MAC aspects and parameter awareness.

Parameter awareMAC protocols

MAC layer aspects

Pape

r

Year

Cha

nnel

Phys

ical

laye

r

Ene

rgy

Mem

ory

Posi

tion/

Dis

tanc

e

Ban

dwid

thad

aptiv

e

Han

dsh

ake

Sync

hron

izat

ion

Nei

ghbo

rdi

scov

ery

Cha

nnel

acce

ssm

etho

d

chan

nel

sele

ctio

n

cari

erse

nsin

g

Sche

dulin

g

Cro

ssla

yer

MA

Cde

sign

Col

lisio

nan

dco

nges

tion

Inte

rfer

ence

Pack

etsi

zean

dst

ruct

ure

Dat

atr

ansm

issi

on

Err

orC

ontr

ol

Del

ayan

dth

roug

hput

Mul

tiple

xing

Bea

mfo

rmin

gan

dsc

anni

ng

Protocol Description

[136]2011 X X X X X X X X X X X X X X X X X X X X X X Effects of congestion and traffic generation intensity are analysed for nano-networks through

competition among bacteria for conjugation at nano gateways.

[137]2012 X X X X X X X X X X X X X X X X X X X X X X The communication and coding schemes are jointly selected to maximise the decoding probability

and minimise the interference while considering energy limitations.

[138]2012 X X X X X X X X X X X X X X X X X X X X X X An energy efficient, scalable and reliable MAC protocol is proposed for dense nanonetworks with

control and data packet structures.

[139]2013 X X X X X X X X X X X X X X X X X X X X X X An energy and spectrum aware MAC protocol is proposed to achieve fair throughput and optimal

channel access by optimising the energy harvesting and consumption in nano-sensors.

[140]2013 X X X X X X X X X X X X X X X X X X X X X X A MAC protocol based on IEEE 802.15.3c is proposed for Terahertz ultra high data rate wireless

networks is proposed with super frame structure and timeslot allocation scheme.

[141]2013 X X X X X X X X X X X X X X X X X X X X X X A MAC protocol is proposed for health monitoring for nanosensor network with anlaysis of node

density and Tx range with routing strategies.

[142]2013 X X X X X X X X X X X X X X X X X X X X X X A MAC protocol for Terahertz communication is proposed with channel access and data rate

analyses with superframe structure.

[143]2014 X X X X X X X X X X X X X X X X X X X X X X A MAC design is proposed for macro scale communication at 100 Gbps for pulse-level beam

switching and energy control with focus on neighbor discovery and scheduling.

[144]2014 X X X X X X X X X X X X X X X X X X X X X X A technique to utilize the harvested energy in wireless nano=networks is presented with focus on

optimal energy consumption for transmission and reception with packet scheduling.

[145]2014 X X X X X X X X X X X X X X X X X X X X X X A frequency hopping scheme is presented to overcome the problems of molecular absorption.

[146]2014 X X X X X X X X X X X X X X X X X X X X X X A receiver initiated MAC protocol is proposed based on distributed and probabilistic schemes for

adaptive energy harvesting nanonodes with scheduling.

[127]2015 X X X X X X X X X X X X X X X X X X X X X X A distributed receiver initiated MAC protocol is proposed with scheduling scheme to minimize

collisions and maximize the utilization of energy harvesting.

[147]2015 X X X X X X X X X X X X X X X X X X X X X X An Rx initiated handshake based link layer synchronization mechanism is proposed to maximise

the channel utilization with analysis of delay, throughput and packet transmission rate.

[148]2015 X X X X X X X X X X X X X X X X X X X X X X A scheme with logical channel information is proposed in which information is encoded in timings

of channel. It supports low rate communication in an energy efficient and reliable manner.

[149]2015 X X X X X X X X X X X X X X X X X X X X X X A cross layer analysis of error control strategeis is presented for nanonetworks with trade-off

between bit error rate, packet error rate, energy consumption and latency.

[150]2015 X X X X X X X X X X X X X X X X X X X X X X An intra-body disease detection is proposed for wireless nanosensor network using on-off keying

and TDMA framework for analysing the data transmission efficiency.

[151]2016 X X X X X X X X X X X X X X X X X X X X X X A fully distributed low-computation scheduling MAC protocol is proposed for maximising network

throughput by jointly considering the energy consumption and harvesting.

[152]2016 X X X X X X X X X X X X X X X X X X X X X X An assisted beam-forming and alignment MAC protocol is proposed with neighbor discovery, data

transmission, delay and throughput analysis.

[153]2016 X X X X X X X X X X X X X X X X X X X X X X A synchronization mechanism is proposed for nano sensor network based on TS-OOK with

analysis of consumed energy, collision probability, delay and throughput.

[154]2016 X X X X X X X X X X X X X X X X X X X X X X A networking approach for static and dense topologies is presented with flooding, network density,

data dissemination and broadcast analysis.

[155]2016 X X X X X X X X X X X X X X X X X X X X X X A link throughput maximization problem is discussed. An optimal packet size is derived with

combined physical and link layer peculiarities.

[156]2016 X X X X X X X X X X X X X X X X X X X X X X Different MAC protocols are compared and analysed in terms of transmission distance, energy

consumption and collision probability.

[106]2016 X X X X X X X X X X X X X X X X X X X X X X Four frequency selection schemes are analysed for throughput and energy utilization.

[157]2016 X X X X X X X X X X X X X X X X X X X X X X A high throughput low delay MAC is proposed to reduces the delay with super-frame structure.

[84]2017 X X X X X X X X X X X X X X X X X X X X X X A high throughput low delay MAC is proposed with on-demand retransmission mechanism based

on verification, reserved timeslots based channel condition and adaptive retransmission mechanism.

[158]2017 X X X X X X X X X X X X X X X X X X X X X X A memory assisted MAC protocol with angular division multiplexing is proposed with multiple

antennas for service discovery, coordination, data communications and interference.

[159]2017 X X X X X X X X X X X X X X X X X X X X X X A hardware processor for 100 Gbps wireless data links is presented. A light weight FEC engine,

BER, frames fragmentation retransmission protocol is also presented.

[160]2017 X X X X X X X X X X X X X X X X X X X X X X A distributed multi radio assisted MAC protocol is proposed with multiple antennas for signal

control mechanism with beam-forming.

[161]2017 X X X X X X X X X X X X X X X X X X X X X X A relay based MAC is presented which considers communication blockage and facing problem.

It further presents a neighbor discovery mechanism and data transmission.

[162]2017 X X X X X X X X X X X X X X X X X X X X X X An adaptive pulse interval scheduling mechanism based on pulse arrival pattern is presented.

[163]2017 X X X X X X X X X X X X X X X X X X X X X X Optimal relaying strategies with cross layer analysis.

[90]2017 X X X X X X X X X X X X X X X X X X X X X X Channel handoff mechanism for mmWave and Terahertz channels, high bandwidth data transfer,

scheduling and channel capacity modelling.

[164]2017 X X X X X X X X X X X X X X X X X X X X X X An energy efficient MAC with clustering and TDMA scheduling for mobility and collisions.

[92]2018 X X X X X X X X X X X X X X X X X X X X X X An autonomous relay algorithm is presented for vehicular networks.

[165]2018 X X X X X X X X X X X X X X X X X X X X X X A secure and intelligent spectrum control strategy is presented with fixed channel.

[55]2018 X X X X X X X X X X X X X X X X X X X X X X Channel switching based on distance, signalling overhead, throughput maximization and error

recovery.

[166]2018 X X X X X X X X X X X X X X X X X X X X X X Throughput maximization with molecular absorption, interference, energy harvesting, and link

capacity.

[167]2018 X X X X X X X X X X X X X X X X X X X X X X Performance of energy consumption with dynamic super-frame durations and packet lengths.

[168]2018 X X X X X X X X X X X X X X X X X X X X X X A MAC Yugi-Ada antenna is presented for frequency and beam direction reconfigurability.

14

band a frequency selective. The spreading loss can be givenas [180],

a1(f, d) =

(c

4πfd

)2

(1)

whereas the absorption loss depends upon the parameterssuch as the temperature, pressure, and distance, and can bedemonstrated as,

a2(f, d) = e−K(f)d (2)

where K(f) is the total absorption coefficient and d is thedistance between transmitter and receiver [181]. K(f) can becalculated using the High-Resolution Transmission MolecularDatabase (HITRAN) database [171].

For a particular transmission distance, the path loss in-creases with frequency due to spreading loss. For a fewmeters distance, the path loss can increase up to 100 dB.Further, the molecular absorption defines several transmissionwindows depending upon the transmission distance. For a fewcentimeters distance, the transmission window behaves likea 10 Terahertz wide transmission window due to negligibleabsorption loss. However, for distance more than 1 meter theabsorption becomes significant which narrows down the trans-mission window. Such extreme path loss results in reducedbandwidth and only a few transmission windows. Differenttransmission windows are marked as feasible in [3] showingup to less than 10 dB of Path loss due to negligible impact ofmolecular absorption. However, due to the spreading loss, thepath loss remains higher, which motivates the usage of highlydirectional antennas and MIMO techniques [3]. The Terahertzwave can be absorbed by raindrops, ices, foliage and grassand any medium containing water molecule [182].

2) Noise: Within the Terahertz band, the molecules pre-sented in the medium are excited by electromagnetic wavesat specific frequencies. These excited molecules vibrate in-ternally where the atom vibrates in periodic motion and themolecule vibrates in a constant translational and rotationalmotion. Due to the internal vibration, the energy of thepropagating waves is converted into kinetic energy partly.From the communication perspective, it can be referred to asa loss of signal. Such molecule vibration at given frequenciescan be obtained by solving the Schrodinger equation forparticular molecular structure [183]. A model for computationof attenuation by gases in the atmosphere is also described byInternational Telecommunication Union, which considers thewater vapor and oxygen molecules over Terahertz band from1-1000 GHz [184]. A HITRAN database is also found usefulfor the computation of attenuation due to molecular absorptionin Terahertz band [171].

The molecular absorption is an important issue to consideralong with free space path loss, as it also causes the loss tothe signals due to partial transformation of electromagneticenergy into internal energy [170], [185]. Such transformationin the Terahertz band can introduce noise which can be dueto atmospheric temperature or the transmission in the radiochannel. The noise occurs due to atmosphere temperature(such as Sun) can be referred as Sky-noise [186]–[188], and

the noise introduced due to transmission in the radio channelcan be referred as the molecular absorption noise [170],[179], [189]. A noise model for Terahertz wireless nanosensornetworks with individual noise sources that impact intra-bodysystems is presented in [190], with noise contributions ofJohnson-Nyquist, black body, and Doppler-shift induced noise.

Molecular absorption noise: The molecular noise is theresult of radiation of absorbed Terahertz energy by moleculeswhich depends on the propagation environment. The funda-mental equation of molecular noise under different assump-tions, such as medium homogeneity or scattering properties,can be directly derived from radiative transfer theory [191].The absorption is generally caused when the transmitted EMwave shifts the medium to higher energy states, where thedifference between the higher and lower energy state of amolecule determines the absorption energy which is drawnfrom the EM wave. It has a direct impact on the frequencyas the absorbed energy is E = hf , where h is the Planck’sconstant and f is frequency [183]. It can also be describedstochastically using the absorption coefficient Ka(f), whichdescribes the average effective area of molecules per unitvolume and depends upon frequency due to which the Tera-hertz band has a unique frequency-selective absorption profile.Similarly, the amount of radiation capable of penetratingthrough the absorption medium is known as transmittance,which can also be defined by the Beer’s Lambert’s Law [170],[185], [191] (2). Further details on the calculation of molecularabsorption coefficient and model can be found in [170], [185],[192], [193].

Sky noise: The Sky-noise is independent of the transmittedsignals and can be known as background noise. It is causedby the temperature of the absorbing atmosphere and can betermed as an effective blackbody radiator or grey body radiatorfor a non-homogeneous atmosphere medium. Several papershave described the Sky-noise, like [186], [188], [191], [194],[195]. It is identified in satellite communication and is mostlyaffected by the antenna temperature which is an additionaltemperature that accounts for the radiation from the absorbingtemperature. The atmosphere can be considered as a dynamicmedium with decreasing temperature and pressure as a func-tion of elevation. In general, it depends upon the absorptioncoefficient and the distance due to the variable temperatureand pressure in the atmosphere [179]. When the distance issmall and the atmosphere is more likely homogeneous theabsorption coefficient can be given as Ka(s, f) = Ka(f)where s represents the distance.

Black body noise: A body with temperature T radiatesenergy, the energy can reach its maximum value for a givenwavelength according to the Wien displacement law [196].This phenomenon is known as black body radiation and itcontributes for a specific range of temperature to the total noiseof the Terahertz system [190], [191].

3) Terahertz scattering and reflection: Reflection and scat-tering are two physical properties that characterize electro-magnetic wave, the region between transmitter and receivercan contain a large number of scatters with different sizesand are distributed randomly. There are two types of scat-tering: elastic scattering in which only the direction of the

15

wave is changed, and inelastic scattering in which the scatterintroduces a change to the energy. The scattering processesinclude Rayleigh scattering which occurs when the dimensionof scatter diameter is larger than the Terahertz wavelengthand Mie scattering otherwise. Mie and Rayleigh scatteringcan affect received Terahertz signal [197]–[199]. In [200], astatistical model for Terahertz scatters channel is proposed,based on indoor measurements of a point-to-point link andtransmitter and receiver were equipped with directional anten-nas, at 300GHz window and a bandwidth equal to 10GHz. Ascattering power radiated from different surfaces is analyzedin [201] for spectrum ranging from 1 GHz to 1 THz. Twoscattering models are analyzed, the direct scattering model andradar cross-section model. The scattering can be an importantpropagation mechanism and for frequencies above 100 GHz,it can be treated as a simple reflection [201].

Radio wave reflections occur commonly in indoor scenarios.The reflected ray depends on the electromagnetic propertiesof the reflector, the surface roughness and the location ofthe reflectors with respect to the transmitter and receiver.The received signal at the RX side is the sum of directray and all reflected rays. In [202], a demonstrator is set upfor four frequency windows: 100, 200, 300 and 400 GHz,to characterize reflections in each window. The reflectioncoefficient is given by:

r =Zcos(θi)− Z0cos(θt)

Zcos(θi) + Z0cos(θt)(3)

where, Z0 = 337Ω is the wave impedance in free space and Zthe impedance of the reflector. Z depends on frequency, ma-terial relative index and absorption coefficient. θi is the angleof incidence and θt is the angle of transmission. The reflectedwave can be reduced using phased array antenna [200], [202].

4) Multi-path: In the presence of reflectors and scatters,a non-line of sight (NLOS) can be generated by the channelfor Terahertz waves. In Terahertz communication, where theline of sight (LOS) and NLOS exist together, the NLOScan interfere with the main signal in LOS at the receivingside [203]. The advantage of NLOS component is when theLOS is obstructed, the receiver can still decode the transmittedsignal. The magnitude of the received signal at the receiverdepends on parameters such as reflector permittivity whichcharacterizes the material, reflector roughness coefficient, in-cidence angle, and wave polarization and finally its positiontoward transmitter and receiver [139], [181]. The magnitudeof the NLOS signal is also affected by the antenna properties,the distance between the source and receiver, and the planecontaining the reflector.

Both LOS and NLOS propagation scenarios exist in theindoor environment [204] where the presence of NLOS ismainly due to scatters and reflectors. The channel attenuationsand delays can be estimated using NLOS and LOS componentsof channel impulse response h(f, t) by:

h(f, t) =√l(d0, f)δ(t−t0)+

NNLOS∑j=1

√l′(dj , f)δ(t−tj) (4)

where, NNLOS is the number of NLOS paths, d is the distance,f is frequency, δ is the Dirac function, and l is the totalattenuation and can be written as,

l(d0, f) = a1(d0, f) ∗ a2(d0, f) (5)

and,

l′(dj , f) = r2a1(dj , f) ∗ a2(dj , f) (6)

Delay parameters in Equation 6 affect some of MAC decisionssuch as modulation and coding selection module, antennabeam steering module, then the estimation of delay parameterscan help to select or switching to the path that gives the lowestattenuation for the link. Presence of NLOS and LOS compo-nents can be used as an alternative for link communicationoutage in which if LOS is blocked the NLOS can be used asan alternate path.

5) Terahertz transmission windows: The path losses whichoccur in Terahertz wave communication give these bands afrequency selective behavior in which some chunks of bandscan be used to provide higher bandwidth due to less amountof losses. Terahertz windows for communication depends onmany parameters, such as communication range and technol-ogy requirements. The distance-dependant bandwidth is givenby [139]:

B3dB(d) = f |a1(f, d)a2(f, d)

N(f, d)≥ a1(f0, d)a2(f0, d)

N(f0, d)−3dB

(7)where, f0 is the central frequency, a1 is the spreading loss, a2is the absorption loss and N(f, d) is the total molecular noise.

Typically, four Terahertz windows can be exploited withinthe band [0.1−1THz] for a communication range of 1−10m.The optimal compact window with low attenuations andhigh bandwidth is the one centered around 0.3THz. The300GHz window is characterized by an available bandwidthof 69.12GHz, subdivided into separate channels or sub-bands.The supported channels for Terahertz communication for thefrequency range from Fmin = 252.72 GHz to Fmax = 321.84GHz were proposed by IEEE 802.15.3d wireless personalarea networks (WPAN) working group and summarized inTable VIII [82], [169]. In [205], transmission windows areselected based on the distance between the nodes becauseof higher attenuation in channel impulse response due tomolecular absorption at longer distance.

B. Design issues and considerations for Terahertz MAC pro-tocols

The Terahertz band can provide high bandwidth for futurehigh-speed networks. However, possesses unique features asdiscussed in the above section which can affect communi-cation performance. These features do not affect the MACperformance directly but impact hugely the Physical layerdesign, antenna and link capacity, which affects the MAC-layer performance, throughput, and latency. The choice ofphysical layer functionalities can also affect the MAC layerdesign such as antenna technology, modulation and coding

16

TABLE VIII: Bandwidth and maximum achievable data ratefor sub bands of 0.3 THz window for single career using highorder modulation [82].

Bandwidth(GHz) Index range Data Rate (Gbps)2.16 1− 32 9.864.32 33− 48 19.718.64 49− 56 39.4212.96 57− 61 59.1417.28 62− 65 78.8525.92 66− 67 118.2751.84 68 236.5469.12 69 315.39

scheme, and waveform. MAC functionalities depends onchannel characteristics, device technology, and physical layerfeatures. There are several design issues related to Physical andMAC layer features which should be considered for designingan efficient MAC protocol for different applications. Theseissues and considerations are highlighted in Table IX withTerahertz features and decisions to be taken at MAC layer.

1) Physical layer and device related issues and considera-tions:

Antenna technology: The transmitted Terahertz signal un-dergoes several impairments due to the propagation througha medium, ranging from free space loss caused by highfrequencies to molecular absorption noise and scattering.To cope with this issue, high gain antennas are requiredto strengthen the signal in one particular direction and tocompensate losses [208]. In communication networks, manynodes try to access the shared channel, it will be challengingto simultaneously serve all of them if we assume antenna isdirectional.

The antenna technology endowed with fast beam switchingcapability can determine the way nodes access the sharedchannel using narrow beams and by assigning each nodeaccess to the channel for a given time slot assigned by theMAC layer [209]. MAC layer should include an antennasteering module to rapidly steer beams toward the receiver.For example, the switching can be performed at the pulse orpacket level. However, the MAC and antennas of differentnodes should be well synchronized in order to reduce errorsand delays. With optimized antenna gain, it is possible toreach high data rate and good signal quality. Massive MIMOantenna is also envisioned for Terahertz applications which canenhance MAC performance by increasing the data throughputand by serving more nodes in the network using spatialmultiplexing techniques [210], where the number of smallantennas can reach 1024. In macro scale communication itis still an open challenge for deep investigations.

In nano communication networks, nodes inter-distance isshort, therefore antenna is assumed to be isotropic and non-complex. Efforts to design terahertz antennas goes back tosome few years, radiation is possible at this frequency bandfor antennas consisting of materials such as InGaAs andgraphene taking advantages from their chemical and electronicproperties [211]–[214]. Antenna dimensions are in the order

of micrometers for THz frequencies. The second issue to beconsidered is materials to be used for antenna design and feed-ing. Some interesting results achieved for the nano-antennaindustry [212], [215], [216], power consumption and radiatedpower, operating band and directivity are the main propertiesof antenna [217]. The mutual coupling is also an importantchallenge to address due to ultra-dense integration of multi-band nanoantenna arrays. A frequency selective approach isproposed in [218], to reduce the coupling effects for multi-band ultra massive MIMO systems.

To mitigate high attenuation and multipath problem, fastbeam switching and steering techniques can be managed byMAC, phased array antennas are the best choice to meet thisrequirement. The phased arrays can improve the link budgetand also increase fairness among users via beam switchingand steering. However the actual status of technology is stilldeveloping towards reducing the switching time and increasingthe number of antenna elements in order to increase gain.Mathematically, the antenna gain for a uniform planar array ata specific angular direction (θ, φ) describes both functionalitiesof beamforming and beam steering and can be calculatedby [219]:

G(θ, φ) = Gmaxsin(Ma(sin(θ)cos(φ)− ν0))

Msin(a(sin(θ)cos(φ)− ν0))

sin(Nb(sin(θ)sin(φ)− ν1))

Nsin(b(sin(θ)sin(φ)− ν1))

(8)

where a and b are two parameters related to vertical andhorizontal separation between antenna elements and also afunction of Terahertz frequency. ν0 and ν1 are two horizontaland vertical steering parameters. MAC layer should selectproperly ν0 and ν1 to establish a communication link witha node, Gmax denotes the maximum antenna gain.

The Terahertz MAC scheduler maps traffic data to eachantenna beam as depicted in Figure 5, the diagram showsan example of mapping between user data traffic and an-tenna beams. MAC selects, based on traffic requirements, foreach transmission time interval [(n − 1)T, nT ], a destinationnode and its associated beam to transmit data traffic. Manyscheduling algorithms can be used such as Round Robin,maximum throughput or minimum delay algorithms. Theswitching operation of the beam can be performed at a pulse,symbol or frame level.

Interference model and SINR: Interference exists in theTerahertz communication system and it can be generated fromnodes using the same frequency band at the same time orfrom the signal itself. Interferences can be caused also byreflected and scattered signals [89], [220] for either fixed ormobile users. Research works on interference modeling arenot sufficient, as a signal to interference ratio depends mainlyon the channel model. The interference level affects signalquality and leads to higher bit error rate (BER). The designof MAC should be aware of interference level by enhancingnodes synchronization or adopting channelization methods toaccess the channel. Access methods define the way each nodetransmits its data, then elaborating an interference model willhelp deeply selecting the right access technique.

17

TABLE IX: Terahertz features, design issues and considered in existing Terahertz MAC protocols.

Terahertz MAC design issues and considerations

THz band fea-tures

Physical layer & devicerelated

MAC layer related

Networkscale

Ref. Application area Path

loss

Noi

se

Scat

teri

ng&

refle

ctio

ns

Mul

tipa

th

Tran

smis

sion

win

dow

s

Cha

nnel

mod

el

Ant

enna

Inte

rfer

ence

and

nois

e

TH

zde

vice

Wav

efor

m

Mod

ulat

ion

and

codi

ng

Lin

kbu

dget

Cha

nnel

capa

city

Cha

nnel

acce

ss

Nei

ghbo

rdi

scov

ery

and

link

esta

blis

hmen

t

Mob

ility

man

agem

ent

and

hand

over

s

Col

lisio

nsan

dm

ulti

user

inte

rfer

ence

Thr

ough

put

and

late

ncy

Rel

iabi

lity

Cov

erag

ean

dco

nnec

tivity

Ene

rgy

cons

ider

atio

ns

nano

[155] Wireless NanosensorNetworks

X X X X X X X X X X X X X X X X X X X X X

[162] Wireless NanosensorNetworks

X X X X X X X X X X X X X X X X X X X X X

[154] Software defined meta-materials

X X X X X X X X X X X X X X X X X X X X X

[206] SDM X X X X X X X X X X X X X X X X X X X X X[149] Nanonetworks X X X X X X X X X X X X X X X X X X X X X[138] Nanonetworks X X X X X X X X X X X X X X X X X X X X X[144] Nanonetworks X X X X X X X X X X X X X X X X X X X X X[146] Nanonetworks X X X X X X X X X X X X X X X X X X X X X[207] Nanonetworks X X X X X X X X X X X X X X X X X X X X X[141] Health monitoring X X X X X X X X X X X X X X X X X X X X X[150] In-body Nanonetworks X X X X X X X X X X X X X X X X X X X X X[167] Health monitoring X X X X X X X X X X X X X X X X X X X X X[145] Industrial monitoring X X X X X X X X X X X X X X X X X X X X X[106] Agriculture monitoring X X X X X X X X X X X X X X X X X X X X X[168] Wireless Nanosensor

NetworksX X X X X X X X X X X X X X X X X X X X X

[161] Nanonetworks X X X X X X X X X X X X X X X X X X X X X[148] Nanonetworks X X X X X X X X X X X X X X X X X X X X X[136] Health monitoring X X X X X X X X X X X X X X X X X X X X X[147] Nano/Macro-networks X X X X X X X X X X X X X X X X X X X X X[156] Wireless Nanosensor

NetworksX X X X X X X X X X X X X X X X X X X X X

[151] Internet of Nano-Things

X X X X X X X X X X X X X X X X X X X X X

[127] Health monitoring X X X X X X X X X X X X X X X X X X X X X[166] Nanonetworks X X X X X X X X X X X X X X X X X X X X X[137] Nanonetworks X X X X X X X X X X X X X X X X X X X X X[139] Wireless Nanosensor

NetworksX X X X X X X X X X X X X X X X X X X X X

macro

[160] THz communicationnetwork

X X X X X X X X X X X X X X X X X X X X X

[84] THz Wireless PersonalArea Networks

X X X X X X X X X X X X X X X X X X X X X

[90] THz Vehicularnetworks and smallcells, SDN

X X X X X X X X X X X X X X X X X X X X X

[159] THz communicationnetwork

X X X X X X X X X X X X X X X X X X X X X

[163] THz communicationnetwork

X X X X X X X X X X X X X X X X X X X X X

[140] THz Wireless PersonalArea Networks

X X X X X X X X X X X X X X X X X X X X X

[152] THz communicationnetwork

X X X X X X X X X X X X X X X X X X X X X

[92] THz Vehicular network X X X X X X X X X X X X X X X X X X X X X[158] THz communication

network, indoornetworks

X X X X X X X X X X X X X X X X X X X X X

[143] THz communicationnetwork, indoornetworks

X X X X X X X X X X X X X X X X X X X X X

Link budget and capacity: A communication link is char-acterized by link budget and system capacity. Link budgetincludes transmitting power, all gains, and losses. Link qualityis good when the link budget value is higher than the receiversignal to noise ratio threshold. This threshold characterizes thereceiver device as well as the bandwidth. Link budget givesthe information of all power gains and losses that should bepresent in one Terahertz link, Terahertz link budget depends onmany factors such as antenna gains, atmospheric attenuations,available bandwidth, distance, temperature, total noise, andthe THz source output power. The link budget should behigher than a fixed threshold which mainly depends on devicetechnology, to guarantee reliable Terahertz communication.Enhancement of link budget leads to increase in reachability,and reduce data loss. It is used as a reference metric to

determine the link range. Shannon capacity is derived from themutual information maximization between sender and receiverfor a particular channel model, it indicates how much data canbe transmitted for a given bandwidth and SINR for differentscenarios.

Most of the studies on Terahertz capacity analysis derivedfor the nanoscale Terahertz network [170], [221], [222], andmacro-scale networks [223] are based on theoretical assump-tions and deterministic propagation channel. In realistic sce-narios, additional properties of the Terahertz wave should beconsidered such as scattering, dispersion, atmospheric factors,and Terahertz statistical model.

Capacity increases with bandwidth and SINR, as a re-sult, MAC layer design should take into consideration theachievable channel capacity for different channel models, as

18

Fig. 5: MAC scheduler with an antenna array. The figure showsthe MAC scheduler module and how it is linked to the antennasystem, if the current node i needs to send data to node j, thenMAC sends a command to the antenna system, coded to steerits beams toward node j using a digital to analog moduleinterfacing between MAC and the physical layer. The beamsteering operation can be repeated for each frame period toschedule a new transmission.

throughput is bounded by the maximum data rate. MAC layershould be aware of link quality: link budget and capacity;frames should be protected against errors, moreover, framelength and transmission duration should be tuned. For MACdesign, link requirements should be considered, for example,in the Datacenter use case, the data rate can exceed 100Gbps,then efficient tracking of link fluctuation is required. Knowl-edge of channel capacity and link budget enhances MACawareness of the channel and the physical layer via frameoptimization and transmission scheduling. In [223], the impactof the outdoor channel on fixed link capacity is studied at a 1km distance. The channel capacity and BER performance fordata transmission are analyzed in [224].

Modulation and coding: The Terahertz channel is char-acterized by high free space loss, molecular absorption, andnoise, as well as limitations of transceivers capabilities. Tomitigate the signal quality issue, modulation and coding arethe main features envisioned. The modulation guarantees anadaptive data rate for fluctuating channel, high order modula-tion for low bit error probability can increase data rate [225],in [226] it is possible to reach 100Gbps using 16-QAMoptical modulation and BER equal to 10−3 and using UTC-PD source for heterodyning; coding helps to reduce errors.MAC layer can be designed to support variable throughputsand fits frame length to channel conditions via informationfrom the physical layer. More works should be addressedto modulation and coding techniques to enhance main MAClayer performance and to design MAC protocol’s physicallayer aware. For nanoscale communication, basic modulationtechniques were used, such as On-Off Keying (OOK) andpulse position modulation (PPM) together with femtosecondpulses [227]–[229]. For improved spectral efficiency, a spatialmodulation technique is proposed and recommended in [230]by using the dense packet arrays of sub-arrays nanoantennaswhile achieving acceptable beamforming performance.

Selection of modulation scheme depends on Terahertzdevice’s capabilities such as output power, bandwidth, andsignal sensitivity. For short-range communication such as nanocommunication, it is possible to use larger bandwidth and alow complexity scheme such as OOK [166] and QPSK [231]can be used. For macro communication, where the range ishigher, the Terahertz gap is split into functional windows andtransmission should be performed using carriers. High ordermodulation can be deployed such as 16-QAM [112] along withthe directional antenna. The channel coding helps to detectand reduce errors at the receiving side, hence reducing thedata loss. However it introduces computational complexity, fornano communication. Therefore, simple coding schemes canbe implemented such as hamming. For macro communicationReed Solomon and Low-Density Parity Checks schemes canbe added [79], [86].

First generation Terahertz devices: Due to high attenuation,devices with good performance such as higher output powerand low noise level are required to optimize the link budgetand increase the link data rate. Terahertz transceivers based onelectronics were developed, for example, Silicon-Germanium(SiGe) based heterojunction bipolar transistor and Gallium-Nitride (GaN) based monolithic mmWave integrated circuit(MMIC) and also transceiver based on photonics such asquantum cascade laser QCL for high frequencies applications.Resonant Tunnelling Diode (RTD) based on InGaAs/AlAsare also promising for Terahertz applications [232], RTDsconvert mm-wave to Terahertz wave. Two kinds of devicesthat perform conversion to terahertz signal: electronic devicessuch as E-RTD and photonic devices such as Uni-TravellingCarrier photo-diodes (UTC-PD). Works on both technologies,photonic and electronic devices, are in progress to choose thebest one of each scenario based on the required data rate,distance, and sensitivity. For some applications, replacing thehigh capacity wire, such as optical link, by Terahertz bridgesis promising as it adds more flexibility to the network andreduces deployment cost. Terahertz devices are responsible forsignal emission and reception. They can affect the link qualitysuch as link budget and the received power spectral density,and can also affect BER and outage probability. Therefore,to design a MAC protocol with a high data rate, devices withlow system noise levels and variable output power are requiredwhile maintaining the SINR in the network. SINR of thenetwork is the result of deviceâAZs transmitted power, channeland system noise. MAC layer, aware of device technologycapability, can monitor power emission, antenna pattern shape,and beam orientation. The antenna technology and deviceperformance can also enhance the node discovery functionalityand reduce delay caused by this operation.

2) MAC layer related issues and considerations:The issues and considerations related to the MAC layer are

discussed below. These issues are also highlighted based onTerahertz applications in Table IX.

Channel access, scheduling, and sharing: For short-rangecoverage such as Nano communication, antenna is assumed tobe isotropic, techniques for random channel access mechanismsuch as CSMA can be then deployed using the same frequencyband. However, techniques to reduce interferences should be

19

implemented as received signals can collide with other signals.For macro-scale communication, it is required to transmit overdistances higher than one meter, antenna should be directionalto overcome channel attenuation effects, different proceduresshould be executed before link establishment such as beamalignment and advanced nodes synchronization. Because, ifthey are not aligned and facing each other, they cannot receivethe transmission, which could introduce the deafness problem.In centralized networks, a central controller is required tomanage the beam alignment with scheduled transmissions andfor Adhoc networks TDMA based approaches can be usedto avoid collisions and manage beam alignment so that eachnode should know when to transmit, to which node to transmitand to which direction. A shared channel can be used amongdifferent nodes, but interference can be increased. An alternateband can be used, which can provide synchronization andcoordination among different nodes to access the channel.

Neighbor discovery and Link establishment: For any com-munication to occur, a link must be established first to discoverthe nodes, which needs to be considered while designing anefficient MAC protocol for nano and macro-scale networks.For nanoscale networks, due to constraints energy storageand generation, mechanisms with low message overhead arerequired. For macro-scale networks, the antenna directionalityand mobility adds extra challenges to establish a link betweenthe nodes, which requires a node to track and locate othernodes for stable link maintenance. For seamless communi-cation, a stable link is required at all times. Efficient beamsteering mechanisms are required to reduce the handshake timeand to reduce the overall neighbor discovery time in Adhocnetworks. Further, due to short-range and mobility, frequentlink association and handovers may be required, which shouldalso be supported in MAC protocol design.

Nodes discovery occurs before the communication phasein which each node needs to inform its neighbors about itsavailability and identity. The discovery phase is constrainedby the lack of information related to node’s position. Fornanonetwork, node can send identity messages to its neighborsand any node receiving a discovery message adds the sender toits list. Generally for dynamic networks, when new nodes canenter or leave the neighboring subset, the discovery procedurecan be more frequent. For static networks, discovery procedurecan be neglected as each node is aware of its neighborsfrom the deployment phase and discovery message can beexchanged after some critical events. For example one ofnodes is out of service or a new node is introduced. Applyingnode discovery in a macro-scale network is challenging ifmobility is added to the system, synchronized beam turningcan be applied during this procedure, again a node receives thediscovery message once beams are aligned, this procedure istime and energy-consuming, optimizations should be appliedto reduce time and energy required to discover neighbors.

Mobility management and handovers: Mobility and cover-age are two mutually correlated concepts, for mobile Terahertzsystem [91], [233] in which radio coverage should be guar-anteed to decrease link outage probability. MAC layer shouldsupport mobility management functionality to guarantee ser-vice continuity. The handover is a technical concept for mobile

networks to describe the changing of the serving base stationwithout interruption of the traffic flow.

One issue rises from nodes mobility is THz localizationdetermination. In [234] authors propose THz RFID techniquesfor device location using specific channel modeling and node’slocalization correlated to the handover procedure. For ex-ample, handover can be triggered in some positions wherereceived power from a serving node is weak. For THz network,localization is more feasible for nanonetworks, however it be-comes more challenging for macro networks where directionalantennas are deployed. Solving the localization problem canhelp to accelerate the handover execution.

Collision avoidance and interference management: Due tohigh bandwidth availability and antenna directionality, it isunlikely that a collision might occur in Terahertz commu-nication. However, it can occur when two node pairs beamdirections crosses each other and perform frequent and longtransmission. The multi-user interference can also occur in ascenario with large number of nodes with mobility. Therefore,a collision detection and avoidance mechanism should be con-sidered while designing an efficient Terahertz MAC protocol.New interference models are required to capture the effectof Terahertz band features and multi user interference [235].The directional communication can decrease the multi-userinterference but it requires tight synchronization between Txand Rx. Further, reduced channel code weights can also resultin lower channel error probability and can also help in avoidingthe multi-user interference and molecular absorption [236].

Reliability: Most of wireless systems require a reliablecommunication, where the degree of reliability defers fromone application to another. The problem becomes more com-plicated when the channel conditions changes with time andcausing time varying absorption [102], [236]. For Terahertzsystems, mainly low frame loss and high throughput arerequired. In Terahertz systems, the eror control module ismainly responsible for frame protection and retransmission toreduce frame losses where frame error depends on channelmodel as well as on frame length. Error control moduleis required especially for harsh channel conditions such asoutdoor channel with dynamic conditions. For nano sensornetwork, a cross optimization method is proposed in [155], toadapt with the frame transmission and size with the channelconditions. Further, for efficient usage reliable wireless linksand beam tracking should be considered in a MAC design.

Throughput and latency: Terahertz band is endowed withlarge bandwidth, due to which it is possible to reach a through-put exceeding 100 Gbps. For some applications like Terahertzdata center scenario, bandwidth is shared between many nodesand therefore a MAC should support and guarantee high datarate and low delay. Fast scheduling algorithms, appropriateMAC techniques and buffering should be implemented to meetapplication specific QoS requirements.

Energy efficiency and harvesting: Energy efficiency meansusing less energy to achieve the required performance. Forsome scenarios it is hard to provide energy in a continuousway such as devices using low capacity battery, mobile nodes.Therefore, energy efficiency becomes a priority for certainTerahertz applications such as nano-communication [144],

20

[237], body area network, fronthaul communication and THzwireless sensor network used in biomedical and militaryfields [238]. To cope with the lack of energy provision,techniques such as energy harvesting, lower modulation ordertechniques can be used to extend the battery lifecycle. It is alsopossible to design new data link layer techniques to increaseenergy efficiency [239]. For applications such as backhauling,Datacenter and information broadcast where the source ofenergy is always available, energy efficiency is less prioritizedthan the previous applications. In Table VI, energy efficiencyrequirement for different applications is highlighted based onthe priority of energy-efficient techniques requirement.

In some applications, the power of nodes is a limitingfactor to transmit continuously, a power management moduleshould be implemented on MAC to reduce power consumptionwithout degradation of the system quality of service. Forexample switching from active to idle state if the node hasno data to transmit and applying the power control strategydepending on channel conditions and target QoS. The secondalternative to save power and guarantee the battery life is toharvest and manage energy, for example for nano-sensors theenergy harvesting is applied for more active nodes. Due tolow storage constraint of nanodevices, the tradeoff must beconsidered for energy harvesting and utilization in an efficientway [10].

Coverage and Connectivity: Terahertz communication ischaracterized by low range connectivity and high availablebandwidth. The coverage or range can be optimized using adirectional antenna, enhanced Terahertz devices, high outputpower, and optimized sensitivity. MAC layer can also con-tribute to coverage and connectivity enhancement by utilizingdata link relaying, path diversity and spectrum switching. Forexample, in vehicular Terahertz network, a mobile node cancoordinated to more than one node. In nanosensor network,for short-range communication, each node can transmit to anynode out of its range using relaying capabilities. Path diversityis also an alternative solution to increase connectivity whena LOS is temporarily unavailable. With no direct LOS linkand to support the seamless communication with less delay,reflectors can also be used to reach out far nodes with no LOSlink [88]. A coverage and achievable rate performance analysisfor multi-user Terahertz with single frequency is presentedin [240].

C. MAC layer decisionsMAC layer is responsible for traffic adaptation with the

physical layer, adding sophisticated modules to optimize linkperformance which can be promising. Following are some ofthe decisions which can be taken at the MAC layer to enhancefurther the system performance.

• Bandwidth and frequency selection: MAC layer shouldactively sense the physical channel as well as be aware ofservice requirements of each traffic flow, then, the selec-tion of the appropriate bandwidth and carrier frequencycan adapt the traffic to the channel as well as reduceinterferences. Most of the actual Terahertz system usesa single frequency, a multiband antenna is also worthconsidering.

• Modulation and coding selection: The Terahertz chan-nel is generally time-dependent, the transmitted signalundergoes impairments leading to high bit error rate, tomitigate this issue, an adaptive modulation, and codingscheme can be adopted and controlled by the MAC layer.High order modulation selection can increase throughputand low order modulation is required to reduce the biterror rate.

• Power management: This module selects the appropriatepower to increase coverage as well as reduce interferenceswhen nodes coordinate between each other. Monitoringnodes using power control can reduce interferences aswell as maintain an acceptable energy consumption value.The power management module, can adapt to its environ-ment, for instance, the mean consumed power value in ahumid environment will be different from a dry one.

• Beam steering: when using a directional antenna forTerahertz communication, beams should be steered ap-propriately to the receiving node, selection of the beamsorientation coordinates can be performed at MAC layerbased on the inputs from the physical layer beam param-eters such as phases between elements.

Implementing the aforementioned modules in the MAClayer, will increase awareness of the physical layer and channelfluctuation as well as to adapt the Terahertz link to theupper layers. Different physical layer functionalities can bemonitored at MAC layer level, for instance, it is possible tochange the modulation scheme from high order 16-QAM tolow order QPSK to reduce bit error rate and from QPSK to 16-QAM to increase data throughput when channel condition isgood, switching operation can be triggered using link qualitystatistics. The module responsible for beam steering can bealso included in the MAC layer, for example using 3 bitsto monitor 8 beams and establish a link with 8 neighbors.Monitoring frequencies and bandwidth can be also included inthe MAC layer, for multi-band wideband antenna, to reduceinterferences and increase data throughput. Finally, the powermanagement module allows monitoring transmitter outputpower to enhance the link budget if the link breakdown orchannel attenuation increases. The power management modulecan take a decision based on collected measurement from thephysical layer and also from other nodes to control signalto interference ratio. Modulation scheme, beams orientation,frequency, and power can be updated at frame level based oncollected statistics from a physical layer as well as reportsfrom the networking layer.

D. Discussion on Terahertz application scenarios

In Table IX, different Terahertz MAC protocols are men-tioned with their application areas. Terahertz band featuresand their MAC design issues and considerations are alsohighlighted. Mainly, these applications are categorized in nanoand macro scale scenarios which are also discussed in SectionIII. Due to unique band features, each MAC protocol ofdifferent applications requires novel MAC mechanisms toaccommodate the high bandwidth availability, path loss, andnoise. Table IX, mentions only those Terahertz applications

21

for which MAC protocol work is available. For nanoscalenetworks, mostly omnidirectional antenna are assumed due tothe short-range and low path loss. For higher transmissionrange the path loss can severely damage the communicationand affect the distance. The Terahertz MAC protocols fornanoscale networks still do not consider the unique featureslike path loss, molecular absorption noise, multipath effect. ForPhysical layer functionalities, the MAC protocols are there asmentioned in the Table IX, but they are not considering theantenna design, channel, propagation, and interference model.

For macro-scale applications, each indoor and outdoorapplication has different requirements and therefore requiredifferent MAC mechanisms. Due to short-range constraint,the Terahertz band suits the indoor applications like TLANand TPAN, these involve mobility with communication overshort distances. The scenarios like Data centers involve staticlinks between different racks and so require point-to-point/-multipoint communications. These scenarios require differentchannel models and scattering and multipath phenomena canaffect communication in a different way. Therefore, whiledesigning an efficient Terahertz MAC protocols, the featuresand design issues mentioned in Table IX should be considered.To enhance the communication range, directional antennashould be used which requires novel mechanisms for beammanagement and tracking with MIMO support, reflectors tomitigate blockage and reach to more than one hop distance.The static points application like KIOSK downloading systemneeds to support quick link establishment and reliability.These applications also required new mechanisms to accessthe Terahertz channel and link establishment, especially whenfrequent link establishment is required and where node densityis high.

The outdoor scenarios like vehicular communication, back-haul, and small cell are interesting scenarios, which involvesmobile and static scenarios. However, the channel can beaffected due to different environmental factors like rain, wind,humidity, and dryness. Therefore, new channel and propa-gation models are required which should also incorporateblocking factors like trees, humans and other physical types ofequipment. Massive MIMO can be used to relay informationbetween cells or nearby networks. Adaptive beam managementcan be utilized by using cooperative massive MIMO andelectronically steerable beams [21]. Further, due to differentenvironmental factors interference mitigation techniques arerequired for outdoor applications.

E. SummaryThe MAC layer protocols play a very important role in

making communication decisions. The environmental andband-specific effects can easily degrade the performance ofTerahertz MAC protocols in terms of delay, throughput, packetreliability, and delivery ratio. Due to unique features of theTerahertz band like noise and path loss, the Terahertz bandcommunication can easily be interrupted compared to theinterference phenomenon in other lower frequency bandslike ISM or GSM. The molecular absorption noise or theatmospheric noise can easily affect the Terahertz commu-nication link and the problem increases with increase in

the distance between the transmitter and receiver. Further,the additional environmental noise factors like Sky-noise canresult in underestimation of noise or interference figure atthe receiver and transmitter which can also affect the MACprotocol performance seriously. The modeling of these factorsis very important in the sense that these factors can behavedifferently in different environments like indoor or outdoorenvironments, and should be modeled carefully depending onthe scenario. Therefore, in designing the MAC protocols forshort, medium or long-range Terahertz communication, theseenvironmental factors, and their modeling must be taken intoaccount. The indoor and outdoor scenarios required differentchannels, propagation and interference models and they needto consider different physical and MAC layer design issuesdiscussed in this section. To strengthen the reflected and scat-tered signals, a metal reflector with goof reflection propertiescan be embedded and to reduce the power absorption thetemperature and humidity can be maintained at a certain levelfor a particular indoor environment. However, for outdoorenvironment, novel mechanisms are required to overcome theeffect of absorption loss.

V. TERAHERTZ MAC PROTOCOLS FOR DIFFERENTNETWORK TOPOLOGIES

In this section, the existing Terahertz MAC protocols arecategorized mainly in network topology as centralized, clus-tered and distributed, as shown in Figure 6. Each topologydesign is then further classified based on the network scale.Different topological designs are considered in the existingliterature based on the application area and its requirementsand are discussed below. In general, the existing TerahertzMAC protocols are characterized and summarized in Table X.

A. Terahertz MAC protocols for Centralized networks

The centralized architecture is mainly followed in nanoscalenetworks due to its limited energy, coverage area, and appli-cation in body area networks [136], [150], [167], [168].

1) Nanoscale networks: The nanonetworks include severalnanodevices that work together to perform simple tasks. Dueto limited energy capacity, centralized topology is used indifferent applications including In-body networks, air qualitymonitoring, and industrial applications. In these applicationsa nano-controller that is capable of performing heavy com-putation, scheduling and transmission tasks is used. Initially,nanodevices send their information to the controller, and con-troller can then process and schedule transmission and sendsinformation to external networks via a gateway device. Theworks in [136], [148], [150], [167], [241] uses a centralizednetwork topology for different nano communication network-related applications. A general figure for such architecture isshown in Figure 7.

A centralized approach is presented in [167] in whichnano nodes can be deployed in a specific area to detectproblems like defects or pollution, where each nano nodecan perform computational tasks with limited memory andtransmits small data over a short-range to the nano-controller.

22

TABLE X: Characterization of existing Terahertz MAC protocols.Year Paper Band Network type Network

scaleTopology Simulator Simulation parameters Analytical

ModelTx/Rx initiatedcommunication

ModulationScheme

Channelaccessmethod

Antenna Complexity

2011 [136] THzbands

Nanonetworks Nano centralized C++ delay, throughput, performance anal-ysis based on distance, propagationtime, packet lifetime

X Transmitterinitiated

- - nano -

2012 [137] 0.1 - 10THz

Nanonetworks Nano distributed custom Energy consumption, Delay,Throughput

X Transmitterinitiated

RD-TS-OOK

TDMA nano -

[138] THzbands

Nanonetworks Nano clustered No X X Transmitterinitiated

- TDMA nano -

2013

[139] 0.1 - 10THz

WirelessNanosensorNetworks

Nano Centralized custom Throughput, optimal channel access,network lifetime, critical transmis-sion ratio

X Transmitterinitiated

pulsebased,TS-OOK

TDMA nano -

[140] 0.1 - 10THz

TerahertzWirelessPersonal AreaNetworks

Macro Distributed OPNet data trasmission rate, avg access de-lay, time for transmitting frames, ac-cess success rate

X Transmitterinitiated

Hybrid - -

[141] THzbands

Nanonetworks Nano Distributed NanoSim- NS3

packet loss ratio, physical transmis-sion

Transmitterinitiated

OOK Random nano -

[142] 0.1 - 10THz

TerahertzCommunica-tion Network

Macro Distributed - - Transmitterinitiated

- Hybrid - -

2014

[143] 0.1 - 10THz

Terahertz net-works

Macro Distributed custom Data rate, throughput X Transmitterinitiated

TS-OOK TDMA directional -

[144] THzbands

Nanonetworks Nano Distributed Matlab energy efficiency, harvest rate, packetbalance

X Receiver initiated OOK TDMA nano -

[145] THzbands

WirelessNanosensorNetworks

Nano Distributed custom SNR, BER, capacity X Transmitterinitiated

pulse based FTDMA - -

[146] 0.1 - 10THz

Nanonetworks Nano Distributed NanoSim- NS3

Prob collision, RTR, fairness, X receiver initiated OOK TDMA nano -

2015

[127] 0.1 - 10THz

Nanonetworks Nano Distributed NanoSim- NS3

Delay, energy consumption, utiliza-tion capacity

X Receiver initiated Pulse basedmodulation

TDMA nano -

[147] 1.04 THz TerahertzCommunica-tion Networks

Macro,Nano

Distributed NS3 Delay, throughput, Packet deliveryratio

X Receiver initiated PSK, TS-OOK

CSMA omni anddirec-tional

-

[148] 0.1 - 10THz

Nanonetworks Nano Distributed custom Failure probability, normalized en-ergy per bot, number of retransmis-sions

X Transmitterinitiated

- TDMA nano -

[150] THzbands

WirelessNanosensorNetworks

Nano Centralized custom Energy consumption, Delay,Throughput

X Transmitterinitiated

OOK TDMA nano -

2016

[149] 0.1 - 10THz

Nanonetworks Nano Distributed COMSOLmulti-physics

BER, PER, enerhgy consumption, la-tency, throughput

X Transmitterinitiated

OOK - nano -

[151] 0.1 - 10THz

Internet ofNano Things

Nano Distributed custom delivery ration, debt, throughput X Transmitterinitiated

- CSMA nano-antenna

-

[152] 2.4 GHz,0.1-10Thz

TerahertzCommunica-tion Networks

Macro Distributed custom packet delay, throughput, failureprobability,

X Transmitterinitiated

- Random,multipleradios

omni anddirec-tional

-

[153] 0.1 - 10THz

WirelessNanosensorNetworks

Nano Distributed Matlab delay, throughput, collision probabil-ity

X Transmitterinitiated

TS-OOK TDMA nano -

[154] 100 GHz Nanonetworks Nano distributed Any-Logicplatform

Coverage, Packet Transmission rate,classification time, collision

X Transmitterinitiated

OOK - nano O(x)

[155] 0.1 - 10THz

WirelessNanosensorNetworks

Nano Distributed COMSOLmulti-physics

link efficiency, optimal packet lengthwith distance

X Transmitterinitiated

OOK - nano-antenna

-

[156] THzbands

WirelessNanosensorNetworks

Nano Distributed Matlab collision probability, energy con-sumption, transmission distance

Transmitterinitiated

OOK - nano -

[106] 1-2 THz Nanonetworks Nano Distributed custom Throughput, Transmission probabil-ity

X Transmitterinitiated

OOK FTDMA nano-antenna

-

[157] 0.1 - 10THz

TerahertzCommunica-tion networks

Macro distributed - - - Transmitterinitiated

- Hybrid - -

2017

[84] 340 GHz TerahertzWirelessPersonal AreaNetworks

Macro Distributed OPNet delay,throughput, success rate, bufferoverflow rate

X Transmitterinitiated

- Hybrid omni -

[158] 0.06 - 10THz

TerahertzCommunica-tion Networks

Macro Centralized custom Throughput, Data rate, Delay, andoutage probability

X Transmitterinitiated

Pulse wave-form modu-lation

Scheduled,multipleradios

omni anddirec-tional

-

[159] 240 GHz Terahertz net-works

Macro Matlab probability of successful frame re-ception, goodput Gbps, percentage oflost headers, ACK frame size, energyper bit

X Transmitterinitiated

PSSS,PAM-16

- - -

[160] 2.4 GHz,0.1-10Thz

TerahertzCommunica-tion Networks

Macro Distributed MonteCarlo

Delay, Throughput, outage probabil-ity

X Transmitterinitiated

Pulse wavemodulation

Random,multipleradios

omni anddirec-tional

-

[161] 2.4 GHz,0.1-10Thz

Nanonetworks Nano Distributed NS3 Throughput X Transmitterinitiated

- Random,mulitpleradios

omni anddirec-tional

-

[162] 100 GHz WirelessNanosensorNetworks

Nano Distributed NS3 Bandwidth efficiency, pulse drop ra-tio, packet deliver ratio, fairness, en-ergy consumption

X Transmitterinitiated

OOK TDMA nano O(N′)

[163] 1.0345THz

TerahertzCommunica-tion Networks

Macro Distributed custom Throughout, optimal distance X Receiver initiated - CSMA directionalantenna

-

[90] 73 GHzand 0.86THz

Vehicular Net-work

Macro Distributed custom data transfer X Transmitterinitiated

- Scheduled,multipleradios

directional O(V k)

[164] THzbands

WirelessNanosensorNetworks

Nano clustering NS3 pkt loss ratio, consumed energy, scal-ability

Transmitterinitiated

- TDMA - -

2018

[92] 0.1 - 10THz

Vehicular Net-work

Macro Distributed custom Channel capacity, PSD, number oflinks

X Transmitterinitiated

- - omni -

[165] 0.1 - 10THz

THz MobileHeterogeneousNetwork

Macro Distributed custom Uniformity, Randomness, HammingCorrelation, throughput, BER

X Transmitterinitiated

FTDMA - -

[55] mmWave,0.1 - 10THz

Vehicular Net-work

Macro Distributed custom data transmission rate X Transmitterinitiated

- Scheduled,multipleradios

directional -

[166] 0.1 - 10THz

Nanonetworks Nano Distributed custom Throughput capacity, interferencepower

X Transmitterinitiated

Pulse basedmodulation(OOK)

TDMA nano-antenna

-

[167] 0.1 - 10THz

WirelessNanosensorNetworks

Nano Centralized custom slot assignment rate, energy con-sumption

X Transmitterinitiated

TS-OOK CSMA nano-antenna

-

[168] 2.3 THz TerahertzCommunica-tion Networks

Macro,Nano

Centralized custom Antenna directivity, antenna con-troller overhead

X Transmitterinitiated

DAMC - omni anddirec-tional

-

2019 [112] 0.25 THz Data Centre Macro Distributed custom FER, BER, Pkt loss, retransmission X Transmitterinitiated

16 QAM - directional -

[241] WirelessNanosensorNetworks

Nano Centralized custom Average delay, remaining energy andtransmitted packets

X Receiver initiated OOK - nano O(N), O(NlogN)

23

THz MAC Protocols for differnt network topologies(Sect.V)

CENTRALIZED(Sect.V-A)

Nanoscale networks(Sect.V-A1)

LL-Modelling [136]CSMA-MAC [167]

DESIGN-WNSN [150]TCN [148]

CEH-TDMA [241]SSA-MAC [242]

Macro scalenetworks

(Sect.V-A2)MA-ADM [158]

IHTLD-MAC [84]MAC-TUDWN [140]

TRPLE [143]HLMAC [157]MAC-TC [142]

CLUSTERED(Sect.V-B)

Nanoscale networks(Sect.V-B1)

ES-Aware [139]EEWNSN [164]

EESR-MAC [138]DYNAMIC-

FH [106]

DISTRIBUTED(Sect.V-C)

Nanoscale networks(Sect.V-C1)

PHLAME [137]DRIH-MAC [127]RIH-MAC [146]

DMDS [151]2-state MAC [166]

G-MAC [153]RBMP [161]APIS [162]MGDI [154]

NS-MAC [156]MDP [144]TSN [145]

SMART-MAC [141]SSA-MAC [242]

Macro scale networks(Sect.V-C2)

Terahertz networks:LL-Synch [147]ISCT [165]TAB-MAC [152]MRA-MAC [160]OPT-RS [163]

Vehicular network:ATLR [92]SDN-CONTR [90]B5G [55]

Fig. 6: MAC layer classification based on network topologies.

The gateway can collect the information and send to the Inter-net. A single hop delay-throughput performance is measuredin [136] for the bio-nano communication network in whichbacteria packets travel towards the nano-gateways followingthe attractant particles emitted by the conjugate nano-gateway.The centralized network topology is also used in [148], [150].A centralized TDMA and energy harvesting based protocol ispresented in [241] in which the nano-controller is responsiblefor channel access and time slot allocation for nano nodes. Tocope with energy consumption nodes are also responsible forenergy harvesting to increase node’s life cycle. A centralizedapproach is then used to free some nodes from performingheavy computational tasks and channel access. Due to smallerdistance, the path loss discussed in Section IV-A1 remainslow. However, the interference from near devices can affectthe transmission opportunities and access to the channel (cf.Section IV-B1). In centralized topology, nodes are within asingle hop distance from the controller node.

Depending on the application requirement, different topolo-gies can be followed in nanonetworks. The works in [136],[148], [150], [167] follows a centralised approach. However,for higher nodes density, multi-hop communication is required.Although, in [242] high node density is considered. But,random arrangement of nodes with mobility is not considered.

2) Macro scale networks: Besides, the nanonetworks, cen-tralized architecture is also used for Terahertz communication

Fig. 7: Nano body communication network.

at larger networks such as macro-scale networks. In [158], acentralized network with Terahertz links is presented whichconsists of an AP and multiple nodes, where the AP coordi-nates and schedules the transmissions among the nodes and isable to communicate directly to each node. The interestingissue in such networks will be synchronization among thenodes, where each node using a directional antenna needs topoint towards an AP, which also increases the interference orcollisions probability from the neighbors. In [158], directionalantennas are used for nodes discovery, initial access, and datatransmissions phase without creating disparity problem [243].

24

The problem arises when the directional beams from theAP and the receiving node have to be well aligned, beamsalignment is time and energy-consuming, a beam managementmodule at MAC layer should be implemented for efficientbeam steering. For centralized macro-scale networks, the APis assumed to be equipped with directional antennas whereasthe other nodes are using the omnidirectional antennas and areassumed to switch to the directional antennas mode after theestablishment of initial discovery which can incur more delays.A beam switching access technique is discussed in [158],beam alignment is performed periodically for the initial accessand transmission period. The centralized topologies are alsopresented in [84], [140], [142], [157] in which a piconetcoordinator is assumed to provide time synchronization in-formation to nearby devices and handles the scheduling andaccess control. A MAC design for macro scale communicationat 100 Gbps is discussed in [143] which considers an indoorpicocell, where an AP communicates with a group of usersusing LOS and directed non-LOS.

An interesting challenge for centralized topology for dis-tances higher than a meter will be to efficiently manage thebeam steering and switching among nodes, which is discussedin Section IV-B1. For applications like TPAN and TLAN, mo-bility adds design challenges for MAC and the channel accessand scheduling can be managed by a central controller (cf.Section IV-B2). In indoor environments with small distancesand mobile nodes, multi-user challenges should be addressedwhich includes efficient mobility and interference managementand resource scheduling.

B. Terahertz MAC protocols for Clustered networks

In a clustered architecture of nodes, a cluster head is electedby nodes in the cluster, the cluster head is responsible for dataprocessing and transmission to the gateway node and to othercluster heads. The clustered architecture is so far seen onlyin nanonetworks environment which also incurs low path lossdue to short distance. Energy efficiency can be improved usingclustered architecture when using the central node.

1) Nanoscale networks: In hierarchical architecture, thenetwork is mainly partitioned in a set of clusters where eachcluster is locally coordinated by a nano controller. The nanocontroller is a device that has more processing capabilities ofcomplex tasks and has high energy availability. Since nanosen-sor nodes are not capable of processing and handling complextasks, these tasks are pushed towards the nano-controllerswhich then coordinate their tasks in an efficient manner. TheMAC layer for nano-controller includes more functionalitiessuch as link establishment and resource allocation. A clusteredarchitecture is followed in [139]. The nanosensor nodes arebattery limited devices, which only can store enough energyto perform few simple tasks. The clustered approach canbe used to manage high node density where inter-clustercommunication can be used to enhance network accessibility.In [139], the transmission and harvesting slots are assignedamong the different nanosensors within each cluster in a waythat the harvested and consumed energy are balanced amongnodes.

A cluster-based nano-network is also discussed for densenetworks in [138], [164], in which inter and intracluster com-munications are carried out to reach gateway node. For plantcommunication, a clustered based architecture is followedin [106] which addresses the frequency selection problemin the Terahertz band which are considered as frequencyselective bands. The nanodevice clusters are used to monitorthe chemical reaction in plants that schedules the transmissionamong themselves and transmit the data to microdevices,which then transmit to the Internet via a gateway device.

The clustered architecture requires efficient MAC protocolsto handle efficient data relaying among the inter and intraclusters. Nanonetworks can utilize very high bandwidth due tolow path losses at small distances with basic modulation andcoding schemes. In such environments, a challenge, related toefficient scheduling and channel access, rises from the factof large number of nodes (cf. Section IV-B2). For higherdata rates, the nano controller can be pushed towards physicallayer synchronization (cf. Section IV-B2). In addition, efficientmechanisms are also required to perform grouping of nanonodes, the dynamic migration of nodes from one cluster toanother, placement and selection of cluster heads constrainedby energy consumption and efficient communication.

C. Terahertz MAC protocols for Distributed networks

Depends upon the application requirements, the devicesboth in the nano and macro scale can perform communica-tion tasks in a distributed manner. The details of the worksfollowing the distributed management are as follows,

1) Nanoscale networks: In distributed network architecture,nodes perform tasks individually and takes independent deci-sions for communication. Decisions can be easy when nodesare aware of the environment and physical layer parametersmentioned in Section IV-B1. For example, when nodes areaware of the channel access and scheduling states of neighbornodes, overall delay can be minimized and throughput can beincreased. In nanonetworks modulation schemes and schedulescan be negotiated among the nodes to perform communica-tion [137]. The scalable distributed networks are discussedin [127], [146].

Due to the absence of controller in Adhoc nanonetworks,nodes need to schedule their transmissions and coordinateto access the channel without generating the interferences.A distributed scheduling mechanism is proposed [162] fornanosensor networks in which every node takes decisionlocally based on its incoming traffic and channel sensingresults. The proposed protocol is shown to be reliable in datadelivery with optimal throughput and addresses the fundamen-tal challenge of limited memory nanodevices. An adaptivepulse interval scheduling scheme is proposed which schedulesthe arrival pattern of pulses transmitted by a nano sink.

In Adhoc network architectures, nodes positions can berandom at times, then nodes relaying can be essential toguarantee connectivity between distant nodes. A set of newfunctionalities should be taken into consideration such asupdating the neighboring list for each node. A multihopnetwork can emerge which needs further investigation [141],

25

[166]. The work in [161] is addressing the antenna facingproblem, in which nanosensors that are in direct range canbe communicated using an omnidirectional antenna, whereasto communicate with other message stations in presence ofobstacles, relay nodes are used with directional antennas.The work is shown to improve the throughput, however,due to the limited capacity of nano nodes, it can increasethe message and energy overhead. A flooding scheme forhigh node density ad-hoc nanonetworks is proposed in [154],in which a message from the external entity is broadcastin a nano-network with coverage in terms of percentage ofreceiver nodes. An internal node can also propagate the datatowards the external entity of a gateway which is movable. Afrequency hopping scheme is modeled in [145] to overcomeattenuation and noise issues using multiple channels betweentwo nanosensor nodes. For distributed networks, the energyproblem is considered in [144], [153], [156].

Different topologies require different solutions based on theapplication requirement, band limitations and design criteria.For example, the small transmission distance and size of nodesrequire the nodes to be close to each other. Depending on theapplication the number of the nodes can be huge. Althoughomnidirectional antennas can be used, high number of nodescan introduce collisions [166], which is not discussed in theseworks. Further, the band limitations and high node densitywith energy-efficient solutions are also not seen in combinedwork.

2) Macro scale networks: Besides nanoscale networks, theTerahertz band promises the ultra-high-speed wireless linksfor macro-scale networks. However, the free space path lossaffects the throughput and results in a reduced coverage area.To extend the coverage and minimize the path losses thedirectional antennas have been encouraged to use. But, will re-quire frequent beam switching and steering. An efficient beamscheduling mechanism can be helpful in providing continuoustransmission with controlled delay. In mobile environments,the Adhoc nature can bring frequent topology changes thatneed to be sorted out using a MAC layer protocol. Forexample, nodes will require mechanism to facilitate nodesentering and leaving the network, relaying, handovers andefficient routes between source and destination with lowestdelay (cf. Section IV-B).

The Terahertz communication network with high-speedTerahertz wireless link is presented in [147] for macro scalecommunication. It addresses the problem of handshaking withantenna speed considerations as an important factor to considerwhile designing a MAC protocol. Nodes are placed within anarea of 10 m circular area in a distributed manner. The mobileheterogeneous architecture is presented in [55], [165] for Ad-hoc connectivity and WLAN to provide high-speed Terahertzlinks and broadband access using access points. In [165], anintelligent and secure spectrum control strategy is proposedfor an indoor network with different access subnets and ananti-jamming strategy with adaptive frequency slot numberselection.

The Omni-directional antennas can be used to performthe initial link establishment and for data transfer directionalantennas can be used to reach further nodes. But, it introduces

some challenges including alignment and synchronization toperform a task and extra antenna overhead. A distributedTerahertz communication network using both directional andomnidirectional antennas is proposed in [152], in which theanchor nodes are used with regular nodes. The anchor nodesare assumed to know their location in advance and regularnodes are equipped with beamforming antenna arrays. Similarwork with Omni and directional antenna is described in [160]where control signals are used for beam alignment using2.4 GHz link and for data transfer Terahertz links are used.Although using 2.4 GHz band for control signaling reduces thehandshaking delay between the nodes, it limits the coveragearea and can leave isolated areas in a network, which furtherrequires multihop strategies to increase the reachability ofthe network. A relaying strategy is proposed in [163] ina network with randomly distributed nodes. However, onlyfew dedicated relays are used to transfer data and nodes areassumed as switching between the transmission and receivingmodes, which can increase the delays.

A software-defined network (SDN) based vehicular networkis considered with distance-dependent spectrum switching,where mmWave and Terahertz band are used alternativelybased on the data transfer usage [55]. It is argued that theuniversal coverage is not possible by using just Terahertzband and therefore a network architecture is proposed, whichuses the microwave, mmWave, and Terahertz bands togetherto achieve the design goal of coverage and channel access.Although, it can extend the coverage, the switching delayincreases as nodes number increases and traffic between nodesincreases. Similar work is presented in [90], which discussesthe handoff and MAC protocol to dynamically switch betweenthe mmWave and Terahertz band for high bandwidth datatransfer operations. Although, performance is shown to beimproved the message overhead and switching delays arehigh. Further, the synchronization is not focussed on multiplevehicles. Another relay algorithm for autonomous vehicularcommunication using Terahertz band links to overcome short-range and unstable links is presented in [92].

The distributed arrangement of nodes can cause discon-nected networks when directional antennas are used. It canalso introduce the deafness problem. Synchronization amongthe nodes to align antenna and exchange neighbor informationwill be another challenge. To solve the synchronization issueworks in [152], [160] uses multiple bands, which can increasethe hardware cost and switching delays. A work presentedin [147], [244] provides link layer synchronization whileconsidering the Terahertz band features. However, the full-beam sweeping time can increase the synchronization delaywhen nodes will be unaware of other nodes and their beamdirection.

D. Summary and discussion

The network topology is an important aspect to consider forTerahertz MAC protocol for which the application scenario,target users, mobility, antenna directionality, and coverageshould be considered. Each application has different topologyrequirements. In nano-communication networks, the nodes are

26

placed at a very small distance from each other. Due to thenear node placement, the path loss is less effective in a nanocommunication network. The omnidirectional antennas can beused in nano-network due to the near placement of nodes.The Omni-directional usage of such scenarios requires a MACprotocol to include collision avoidance methods with efficientsensing mechanism to detect interferences. The path lossincreases with distance, therefore to mitigate free space attenu-ation effect, directional antennas are required. The antenna di-rectionality requirement clearly impacts on link establishmentand channel access mechanisms. The transmission schedulecan be easily managed in a centralized scenario, in whicha central controller is responsible for overall transmissionschedules which also requires energy-efficient mechanisms.However, in distributed networks, scheduling the transmissionsand resources is a challenge, especially when directionalantennas are in use over the short distance.

At macro scale, for Terahertz communication networks,the topology must account for many practical concerns likescalability, reconfigurability, LOS connectivity due to an-tenna directionality requirement, fault tolerance, and cost-performance index. In indoor scenarios the distance betweenthe APs and users with mobility support should be coveredin the MAC protocol design while providing fault-free andseamless communication. The distributed topology for Tera-hertz MAC protocol must accommodate the dynamic nature ofthe network while covering the whole network. The TerahertzDatacentre network, for example, requires top of rack nodesto transmit data among different racks. The short-range limitsthe connectivity of the nodes, therefore, novel mechanisms arerequired to approach far distance nodes within a Data Centre.

VI. CHANNEL ACCESS MECHANISM FOR TERAHERTZCOMMUNICATIONS

In this section, the existing channel access mechanismsfor Terahertz band communications are presented. They aremainly classified based on macro and nanoscale Terahertznetwork. They are further classified as Random, Scheduledand Hybrid channel access mechanisms as shown in Figure 8and discussed below.

A. Nanoscale Networks

The pulse-based communication in nanonetwork transfersinformation using short pulses which reduces the chanceof having collisions (cf. Section IV-B1). To avoid possiblecollision, the duration between two pulses can be increased toallow different users stream at the same time. A nanodevicecan send a packet when it has something to send withoutwaiting in a random way, where receiver devices should beable to detect such pulse. A node can also be aware of orpredict the next transmission from the received packet.

In nanonetworks several nanosensor nodes, taking randompositions, can be used to maintain the network connectivityfor different applications, such as in-body sensing, toxic gasdetection and control, and military fields. The sensing andcommunication capabilities limit their target area to few mil-limeters. Although, smaller size antennas can be integrated

and massive data can be exchanged at high data rates. Thisrequires a simple, robust and energy-efficient channel accessmechanism for communication among the nano nodes. Differ-ent existing channel access mechanisms for nanonetworks arediscussed below and shown in Figure 8.

1) Random channel access: In Random-access mecha-nisms, different nodes contend for the channel access ortransmit packet in a random manner. The random access isnot suitable for applications in which higher number of nanonodes are used. This is due to the limited sensing, computation,battery and memory capacity of nanodevices, which can allowonly a few transmissions until another harvesting phase isrequired.

The carrier-based channel access mechanisms are mostlyunsuitable for nano communication due to extra sensing over-head and energy consumption. In [167], a slotted CSMA/CA-based channel access mechanism (CSMA-MAC) is proposedin which nodes contend for the channel. An energy harvestingmodel is also presented. The slots usage is found higherwhen slotted CSMA method is used. In addition, the super-frame duration and packet size are also mentioned to beconsidered for slotted CSMA protocols. In these types ofnetworks, a beacon can be used to either synchronize thenext transmissions or to ask for transmission of data packetsdirectly. In direct beacon transmission for data transmission,collision can occur as two nodes can transmit at the same timeand it also needs to address the energy constraint as frequentpacket transmission can drain the nanodevice energy fairlyquickly. Another work that uses carrier sensing is presentedin [151], where sensing duration is used to optimize thetransmission schedules.

An simple Aloha based channel access mechanism is pre-sented in [141], as Smart MAC in which nodes performhandshake before sending a packet to know about one-hopneighbors. When there will be no neighbor to transmit the nodewill apply a random backoff delay prior to start again anotherhandshake mechanism. The receiving node verifies when thereare any physical collisions. The collisions can occur whenthere is a higher number of nodes available, which is notaddressed and the retransmission mechanism is not discussed.

A random channel access with multiple radios is used fornanonetworks in [161]. For control signal transmission 2.4GHz band is used and for data transmission Terahertz band.The channel is accessed in a random manner in both phases.It also addresses the antenna facing problem in first phaseby synchronizing the antenna directions for the data trans-mission phases. Because the narrow beams can not cover thewhole search space. It mainly overcomes the synchronizationproblem using directions antennas as in Terahertz band at thecost of multiple radios. The alignment between two phases isrequired to decide the time in each phase of transmission.

When the number of nodes is higher, random access orpacket transmission can pose several challenges including col-lisions, transmission delays, and higher energy consumption.However, the energy-efficient harvesting mechanisms are notconsidered in these works. The works in [141], [167] donot consider the limited memory and energy limitations of

27

Channel Access Mechanisms in THz MAC protcols(Sect.VI)

Nanocale Networks(Sect.VI-A)

Random ChannelAccess Mechanisms

(Sect.VI-A1)CSMA:

CSMA-MAC [167]SMART-MAC [141]

DMDS [151]Multiple radios:

RBMP [161]

Scheduled ChannelAccess Mechanisms

(Sect.VI-A2)TDMA:

DESIGN-WNSN [150]TCN [148]

ES-Aware [139]EEWNSN [164]

EESR-MAC [138]PHLAME [137]

DRIH-MAC [127]RIH-MAC [146]

2-state MAC [166]G-MAC [153]

APIS [162]MDP [144]

SSA-MAC [242]FTDMA:

DYNAMIC-FH [106]TSN [145]

Macro Scale Networks(Sect.VI-B)

Random ChannelAccess Mechanisms

(Sect.VI-B1)CSMA/CA based:

LL-Synch [147]OPT-RS [163]

Multiple radios:TAB-MAC [152],MRA-MAC [160],

Sceduled ChannelAccess Mechanisms

(Sect.VI-B2)FTDMA:

ISCT [165]TDMA: TRPLE [143]

Multiple radios:SDN-CONTR [90]

B5G [55],MA-ADM [158]

Hybrid ChannelAccess Mechanisms

(Sect.VI-B3)CSMA/CA

and TDMA:HLMAC [157]

IHTLD-MAC [84]MAC-TUDWN [140]

MAC-TC [142]

Fig. 8: Terahertz Channel Access Mechanisms classification.

nano nodes. In [161], synchronization among the nodes isaddressed but at the cost of multiple radios which increasesthe challenges. The random access mechanisms can increasethe throughput but high message overhead and sensing requiresufficient energy availability. Whereas, the scheduled mecha-nisms, solve the energy issue but at the cost of throughput.A work in [151] provides the optimal schedules with randomsensing at the beginning of the slot for distributed environment.It considers the limited memory and energy consumption ofnano nodes, but not maximizing the overall throughput. Thework is for distributed environment but only few nano nodeswere used for implementation.

2) Scheduled channel access: Nanodevices require a simplecommunication and medium access mechanism to effectivelycollect data from other nanosensor devices. Due to presenceof large number of nanodevices, providing optimal schedulesfor nano nodes life long communication is a challengingtask. In centralized network a nano controller is responsiblemainly for single-hop nodes schedule. However, in distributednetworks, managing schedules for channel access for morethan hop distance nodes is a challenging task. The optimalsolution should also consider the limited nano nodes capacity,huge bandwidth availability and energy fluctuations whiledesigning an efficient and optimal channel access schedulingmechanism [148], [245] (cf. Section IV-B).

TDMA based: In TDMA based approaches, each nodeis assigned a time period to transmit its data. A schedulingmechanism based on a TDMA based approach is presentedin [139], [148], [150] in which a nano-controller makes thedecision for a nanosensor node to transmit the sensing data.In [148] in which a timing-based logical channel concept isused. In these logical channels information can be encoded inthe silence period between two events. Although, these logicalchannels are shown to achieve synchronization among thenodes with energy efficiency, low rate, and collision avoidance.The unique features of Terahertz band are not considered (cf.Section IV-A).

A dynamic scheduling scheme based on TDMA is presentedin [139] as ES-aware, in which a nanosensor dynamicallyassigns variable length transmission time slots which dependupon amount of data to be transmitted; the distance be-tween the nanosensor and controller; and the energy of thenanosensor. To balance the trade-off between the throughputand a lifetime, an optimal scheduling strategy is proposedwhich aims to provide an optimal transmission order forthe nanosensors to maximize the throughput. This algorithmutilizes the inter-symbol spacing for the pulse-based physicallayer to allow a large number of nanosensors to transmits thepackets in parallel without introducing any collisions. Thisworks are mainly for single-hop networks.

28

The TDMA based scheduling for multihop WNSN MAC(EEWNSN) protocol is presented in [164], which takes bene-fits of the clustering techniques to alleviate the mobility effectsand transmission collision. After selecting a nano router, thenano router allocates the specific time slots to the nano nodeaccording to a systematic allocation pattern. The timeslots areconsidered fixed and due to the transmission to the closestnano router, the energy consumption can be decreased whichcan prolong the network lifetime. Another work following acluster-based architecture for nanonetworks (EESR-MAC) isgiven in [138], in which initially a master node is selectedwhich then allocates the transmission schedules between theinter/intra-cluster using TDMA approach. The master node’sroles are periodically changed among different nodes to avoidlong-distance transmissions and to save energy.

In [139], spectrum and energy parameters are considered forschedule assignment. In [137], nodes select different Physicallayer parameters, energy, and channel conditions, agreed byusing a handshake process that can limit the performancedue to the limited capacity of nano nodes. Although, param-eters can be negotiated the dynamic and optimal parameterselection is still required for these networks to increase thelifetime and performance. A Rate division Time-Spread On-Off Keying (RD TS-OOK) is also proposed which is basedon asynchronous exchange of femtosecond long pulses spreadover time. To minimize the probability of multiple sequentialsymbol collisions in a packet, the time between the symbol Tsand symbol rate β = Ts/Tp are chosen differently for differentnanodevices for different packets. When all nanodevices aretransmitting at the same symbol rate, a catastrophic collisioncan occur which can cause collision for all symbols in apacket. The orthogonal time hopping sequences can be used toavoid this condition [246]. The symbol collisions are unlikelyto occur due to a very short length of the transmitted symbolsTp and because the time between symbols T − S is muchlonger than the symbol duration Tp. By allowing differentnanodevices to transmit at different symbol rates, collision in agiven symbol does not lead to multiple consecutive collisionsin the same packet. As an example, an RD TS-OOK illustra-tion is shown in Figure 9 in which two nanodevices transmit toa common receiver with different initial transmission times asτ1 and τ2. A short pulse represents a logical 1 and a silencerepresents a logical 0. The device 1 plot shown a sequence“10100" and device 2 plot shows a sequence of “11100".

A scheduling mechanism for distributed networks is pre-sented in [127] (as DRIH-MAC) using edge colors. Its central-ized version is presented in [146] using probabilistic method.The edge coloring problem is considerably challenging inad-hoc based networks due to the absence of a centralizedcoordinator. In DRIH-MAC [127], the medium access controlrelies on the receiver-initiated and distributed scheduling fornano nodes in which each pair of nano nodes within a commu-nication range will have an edge with different color. The mainobjective is to determine a minimum number of colors requiredto color the edges of a graph i.e., two edges incident on acommon node do not have the same color, where each colorrepresents a timeslot in which a nano node can communicatewith one of its neighbors. At most (δ+1) timeslots are needed

at least to reach an agreement/disagreement on color withall neighbors through RTR packet assuming no RTR packetfailure. These works are shown to be efficient, however, thelimited memory capacity of nano nodes is not considered inthese works.

Due to limited battery capacity, frequent transmission cannot occur using a nanodevice. While designing a channelaccess mechanism, energy harvesting must be considered toachieve optimal network performance. A two-state MAC isproposed in [166] in which two states are used for a nodea harvesting only and harvesting with transmission. Anotherwork using harvesting in sleeping and transmission modes ispresented in [153] for grid-based nanonetworks.

For accommodating bursty traffic in a distributed environ-ment, an adaptive pulse interval scheduling (APIS) schemeis presented in [162]. In this scheme, the arrival patternof pulses transmitted is scheduled by nano sinks based onthe access bandwidth. It has two scheduling steps such astransmission shifting and interleaving which are based oninformation collected from short channel sensing. When nanosinks start transmitting pulses, they are first shifted in sequencewithin interval of IS . After which multi-user transmissions areinterleaved by separating pulses with the interval that evenlyshares the bandwidth among nano sinks and in response thepulses arrive at the gateway in an ideal pattern.

The main issues while designing a scheduled access mech-anism for nanonetworks are to consider the energy harvestingand consumption, limited computational capacity, Terahertzband features and optimal performance with node density. Theworks in [137], [144], [148], [153], [166] considers energyproblem, however, they lack in considering other aspects ofmemory, collisions, and Terahertz band features. In [137],channel parameters, and coding scheme aware protocol ispresented but it does not consider the high node density,balance in energy harvesting and consumption and limitedcomputational capacity of nanodevices. The work is requiredfor distributed nanonetworks while considering the uniqueaspects of Terahertz band. In [164], multihop protocol isdiscussed but energy efficiency is not considered. The nanonodes require an efficient mechanism to balance betweenenergy harvesting and consumption. The work in [242], isshown to be better in performance than [137], [146], [164],and considers self slot allocation with energy for both cen-tralized and distributed environment, but band features are notconsidered and the trade-off between energy harvesting andconsumption is not discussed. In [139], the access scheme isprovided while considering this trade-off with fair throughputand optimal lifetime, where nodes are aware of energy andspectrum information. In this work, the high node density isalso considered for performance evaluation.

Frequency and Time Division Multiple Access (FTDMA)based: In [106], dynamic frequency selection strategy(DYNAMIC-FH) is presented which uses FTDMA. AnFTDMA is initially considered and for a higher numberof nano nodes multi-frequency is proposed with timeslotsscheduling for a different number of users. Each node isassigned with different timeslots to avoid collisions in caseof higher packet sizes (like multimedia traffic). The main

29

Fig. 9: An example of Rate Division Time-Spread channel access mechanism used for pulse based communication in TerhertzNano communication networks [137].

objective of the frequency selection strategies is to minimizeenergy consumption and increase channel capacity. In [145],also a Markov Decision Process-based frequency hoppingscheme (TSN) is proposed in which entire band is dividedinto K frequency sub-channels where the aim is to determinefor each timeslot the subchannels be used.

B. Macro Scale Networks

Figure 8, shows the classification of Terahertz band channelaccess mechanisms for Terahertz macroscale networks. Thesummary and details of each category are discussed below.

1) Random channel access: The examples of randommechanisms are ALOHA and CSMA techniques. Ideally, forthe random mechanism, a node should sense a medium beforeaccessing it. Since there is a large bandwidth available, thechances of collision occurrence are less. Therefore, the randommechanism is also being used followed by message confirma-tion strategies. However, the idea of collision and interferencecannot be ignored completely as there could be many usersaccessing the same medium and might be transferring a largevolume of data which could potentially generate collisionsamong the two nodes (cf. Section IV-B2). The collisionsavoidance schemes and recovery from collision schemes arevery essential. The delay, on the other hand, can be minimizeddue to random access, however, further research is required forthe collisions and delay parameters trade-off.

CSMA based: In carrier sensing based channel accessschemes, control packets are exchanged before data trans-mission to access channel. The high message overhead anddirectional antennas can create problems. To reduce the mes-sage overhead, in [147], a one-way handshake based channelaccess scheme is proposed. A node in transmission modelistens for messages from other nodes until one is received. Asthe directional antennas are used, the antenna facing problemcan occur, however, it is assumed that the nodes know eachother’s position. To address the antenna facing problem, thework is extended in [163] by focusing on the use of highlydirectional antennas to overcome high path loss. In [163], therelaying distance is studied that maximizes the throughput by

considering the cross-layer effects between the channel, theantenna and the communication layers. This work also focuseson control message exchange to establish nodes associationand follows the random channel access, as in [147]. Although,these schemes are shown as workable, but are not consideringthe high node density, message overhead and unique Terahertzband features discussed in Section IV-A.

CSMA with multiple radios or hybrid system: The limitedpower of transceiver and high path loss limits the transmissiondistance as discussed in Section IV-A and can also increase thechannel access problem. Therefore, requires directional anten-nas with beamforming both at the transmission and reception.The beams need to aligned to establish a link and to access achannel before transmitting data packets (cf. Section IV-B1).The beam alignment takes time when antenna sweeps todiscover the neighbors. Therefore, in some works like [152],[160], multiple radios are used to divide initial access anddata transmissions. These works can increase the messageoverhead, increase antenna switching delay at higher radiocosts. The channel is accessed by sending an request/clear-to-send (RTS/CTS) packets including node positions. However,if nodes are mobile, it will cause repeated discovery phase.In [160]. instead of sending a clear to send packet to thetransmitter, the receiver estimates the angle of arrival and sendsTTS packet to the transmitter. The transmitter can then switchand adjust its directional antenna and starts pointing towardsthe receiver antenna to start the data transmission. Althoughthe message overhead is shown to be reduced, the uncertaintyof packet loss during user association phase is not considered,the difference is shown in Figure 10 in which one scheme uses4 transmissions until antenna direction alignment and otheruses two transmissions. The high path and absorption loss arealso not discussed in these works, which can cause a packetloss.

Works in [147], [163] uses random channel access with aone-way handshake to reduce the control message overhead.However, the high node density can cause collisions problemswhich are not considered in these works. Further, synchroniza-tion can be a problem when using directional antennas in theseworks. To solve the synchronization issues, the works in [152],

30

[160], use multi-band antennas with lower frequency bandsand Terahertz bands. Although synchronization problem canbe solved, the antenna switching and alignment can increasethe delay.

Fig. 10: Random channel access, node association with an-tenna direction alignment difference of TAB-MAC [152] andMRA-MAC [160] protocols.

2) Scheduled channel access: In distributed networks as-signing the schedules to Terahertz nodes is an NP-hardproblem [90]. Until the device technology is more matureto allow non-LOS communication, scheduled channel accesscan enhance the network performance in the presence of theabove-mentioned constraints. In scheduled access, each nodeis assigned with a particular timeslot. Some of the variationsof scheduled access are discussed below like FTDMA andTDMA based channel access mechanisms in Terahertz.

FTDMA: An FTDMA based technique is used in [165] inwhich the available frequency is further divided into sub-bandsand assigned frequency slot numbers. In FTDMA, frequencyis divided into different timeslots. The frequency used by anyuser in a particular time could be represented by a sequenceSk. To avoid jamming different sequences or transmissionstrategies are adopted by each user. For n users transmittingat the same time, the sequences used by each user mustbe orthogonal. The performance is shown to be improvedfor security and throughput, however, the unique features ofTerahertz bands are not considered like path loss and noise.

TDMA: A TDMA based channel access scheme is usedin [143] to avoid fighting for access. In [143], a fully direc-tional MAC protocol for the Terahertz network is presentedwhich relies on pulse-level beam-switching and energy control.A MAC frame structure is presented which is composed of aPOLL period, a Downlink (DL) period and an Uplink (UL)period. In the POLL period, the AP learns the traffic demandsof the users and schedules the DL/UL transmissions. In DLand UL, each different user is assigned a separate timeslot toaccess the channel.

TDMA with Multiple radios: In Terahertz band, directionalnarrow beams are required to enhance the transmission dis-tance, as discussed in Section IV-B1, which can increase thedelay to establish initial access, handovers and beam track-ing. The TDMA based approaches suits to assign schedules

for beam alignment and channel access which also requiressynchronization among nodes [55], [90]. A 2.4 GHz bandand mmWave with non-LOS are being used to achieve initialsynchronization and beam alignment, and a Terahertz band canbe used to transfer data. In this way, next time when a nodeenters into a communication range or requires data, beams canbe aligned in advance to perform seamless communication.

A TDMA based channel access scheme is used in [90],in which an SDN based controller (SDNC) is used to switchbetween mmWave and Terahertz band for vehicular communi-cation for high bandwidth data transfer operation. An optimalprocedure at the SDN controller for scheduling multiple vehi-cles for accessing a given small cell tower is also given usinga time division approach. The objective is to maximize thebits exchange between the cell tower and vehicle where thecondition is to at least schedule one car in each timeslot whileconsidering the distance.

For link switching, it is proposed that the Terahertz bandshould be switched whenever the link between the vehiclesand the cell tower is less than dth, and to mmWave other-wise. Another work discussing the hybrid usage by switchingbetween the mmWave, µW band, and Terahertz band isgiven in [55]. In this work, a higher capacity link like theTerahertz band is used for data transfer and mmWave forACK transfer. For error recovery stop and wait is followedwith data transmission using Terahertz band and ACKs usingthe mmWave band. However, this alternate band usage canintroduce excessive overhead, higher delay for receiving anACK, and also introduces the beamforming overhead as thecommunication must be directional for the Terahertz band.

A memory assisted angular division multiplexing MACprotocol (MA-ADM) is proposed in [158] for centralizedarchitectures in which an access point is responsible forthe coordination and scheduling the transmissions to achievefairness and efficiency. It also uses Omni-directional antennasto overcome beam alignment and discovery problems and usesdirectional antennas for data transmission. Memory-Guidedmessage transmission is used by the AP, in which during thenetwork association phase, a node establishes the connectionwith the AP using an angular slot and register it in the memory.The AP switches the narrow beam by checking the memorytowards the registered angular slots for data transmissions toavoid the empty scanning of the unregistered angular slots.The initial use of omnidirectional antenna can limit the servicerange which can affect the connection of nodes with AP, whichis not considered. In addition, the switching delay and cost arealso analyzed. To maintain the fairness in which transmissioncompletion is verified by the reception of an ACK messageor repeated failure occurrence the service discovery phase istriggered to update the guided transmissions. The schedulingalthough is considered for data transmission but not focusedin detail.

In macro-scale network synchronization with directionalnarrow beams is a challenge. The scheduled access can providecontention-free network but increases the delay. A synchro-nization with narrow beams directional antennas can be usedbut beam alignment requires more focus to reduce the switch-ing and transmission delay. Although in works like [143],

31

TDMA approach is followed for channel access, efficientsynchronization is still required. To solve the synchronizationproblem, multiple bands are used in [55], [90] but overheadis high and specific Terahertz band features (cf. Section IV-A)are not considered. A memory assisted approach can be useful,as proposed in [158], for using the beam direction from thememory. It is not considering the Terahertz band features.

3) Hybrid channel access mechanism: The random accessmechanisms can increase the delay and can cause collisionswhen node density is high but can enhance throughput. How-ever, the scheduled access schemes can reduce the impactof collisions but can enhance the throughput performance.Therefore, hybrid mechanisms are required to overcome thelimitations of these schemes. A hybrid channel access mecha-nism is proposed in [84], [140], [142]. In hybrid mechanismsCSMA and TDMA are used in which the channel time isdivided into multiple superframes. Each superframe consistsbeacon period, channel time allocation period (CTAP) andchannel access period (CAP). The CSMA/CA is used tocompete for the channel in CAP period in which the devicewhich wants to transmit data need to send a channel timerequest command to PNC. The PNC broadcast slot assignmentinformation in the net beacon frame according to request framereceived. The devices can synchronize themselves based on thesynchronization information and obtain CTAP slot allocationinformation, which is made of channel time allocation. Thedevices access channel in TDMA mode where each devicetransmits data in its allocated slot. An on-demand retrans-mission mechanism is also proposed to decrease the messageoverhead with reserved slots mechanism based on channelcondition. This work is an extension of [157] which also usesa hybrid system and describes a high throughput and low delayTerahertz MAC protocol. The throughput is also shown to beimproved by updating the timeslot request numbers with areduction in latency efficiency.

The hybrid channel access mechanisms can improve thelimitations of both random and scheduled channel accessschemes. The schemes in [84], [140], [142] can improve theperformance, the poor network conditions are not consid-ered. The new scheme should consider the Terahertz bandfeatures (cf. Section IV-A) and design considerations (cf.Section IV-B).

C. Summary and discussion

The traditional channel access schemes are based on con-tinuous signals which cannot be used for nanonetworks due tothe size and energy constraint. Instead, short pulses (100 fs)can be generated using simple devices (graphene antenna) andtransmitted at the nanoscale. Therefore, novel channel accessmechanisms are required for nanoscale networks. Mostly,scheduling based channel access mechanisms are used to avoidfrequent messaging to contend for channel access and energyconsumption. The new mechanisms should consider the highnetwork density and limited energy availability and in somecases nodes mobilities. Using short pulses can reduce thecollision probability and therefore for high-density networkrandom channel access mechanisms can be used by increasing

the duration between two transmissions. But, in their MAC thecollision avoidance mechanism should be considered to avoidany possible collision and to reduce the message overhead. Byusing isotropic antenna, a node at nanoscale can communicatewith more nodes in its communication range. At the macroscale, the directional antennas are preferred due to high pathloss and coverage enhancement but require antenna directionalignment and beam management. To solve the problem ofantenna alignment different mechanisms are proposed in theexisting literature which includes searching the space andadding an additional tuning phase [82]. But, these mechanismscan increase the link establishment delay and energy. It cancause a hidden node or deafness problem (cf. Sect.IV-B2), inwhich a node remains unaware of the existence of the nearbynode due to limited coverage or antenna misalignment.

Pulse based communication is mainly used in nanonetworkswith channel access schedules using a nano controller. Inpulse-based communication, short Terahertz pulses can begenerated using specific devices to reach higher throughput.Whereas, time division technique for Terahertz MAC is mostlyrelated to the fact that each node has equal share for channelaccess and to avoid possible collision. However, the TDMAbased approaches require efficient and optimal schedulingschemes while considering antenna direction, beam alignment,multi-user interference, and Terahertz band features. To avoidtight synchronization requirement, asynchronous MAC pro-tocols requires further research with energy efficiency. Thechannel access for macro scale relies more on beam steering.Therefore, an alignment phase should be considered at theMAC layer. The antenna pattern for nanoscale is assumed asisotropic and directional for macro-scale network for which thegain requirements are also high. Further, the high interferencefor nanoscale due to isotropic properties of antenna requiresefficient algorithm while considering the channel character-istics, antenna, gain and transmission power. New realisticmeasurement studies are also required for modeling of channelbehavior for different indoor and outdoor applications at macroscale.

Actual terahertz network is still suffering coverage issues,advances in antenna technologies, transmitter and receivertechnologies will enable new medium access and fast schedul-ing techniques and improve network performances.

VII. TRANSMITTER AND RECEIVER INITIATEDTERAHERTZ MAC PROTOCOLS

The existing Terahertz band scenarios like nanoscale andmacro scale networks require different MAC mechanisms.The nanoscale networks are energy-constrained networks inwhich nanodevices have just enough energy to transmit apacket and therefore use energy harvesting mechanisms togenerate energy [10]. In macro-scale networks, directionalityof antenna requires attention in establishing communication.The directional antennas reduce the multiuser interference butit requires tight synchronization between the transmitter andreceiver to overcome hidden node or deafness problem [247].

Establishing a link is an essential part to obtain beforestarting a communication that can be used to achieve syn-chronization and exchange information to establish a network

32

Transmitter and Receiver Initiated THz MAC Protocols(Sect.VII)

Transmitter Initiated MAC protocols(Sect.VII-A)

Nanoscale networks(Sect.VII-A1)

PHLAME [137]ES-AWARE [139]

TCN [148]LL-Modelling [136]

DMDS [151]2-state MAC [166]CSMA-MAC [167]

G-MAC [153]APIS [162]

ERR-CONTROL [149]MGDI [154]

PS-OPT [155]NS-MAC [156]

DYNAMIC-FH [106]EESR-MAC [138]

RBMP [161]SSA-MAC [242]

Macro scale networks(Sect.VII-A2)

IHTLD-MAC [84]ATLR [92]

TRPLE [143]SDNC [90]B5G [55]

MA-ADM [158]TAB-MAC [152]MRA-MAC [160]MAC-YUGI [168]

Receiver Initiated MAC Protocols(Sect.VII-B)

Nanoscale networks(Sect.VII-B1)

DRIH-MAC [127]MDP [144]

RIH-MAC [146]LLsynch [147]

Macro scale networks(Sect.VII-B1)OPT-RS [163]LLsynch [147]

Fig. 11: Classifications of Terahertz MAC protocols based on Transmitter and Receiver initiated communications.

like neighbor information, physical parameters and beamalignment (cf. Section IV-B2). The handshake mechanismsshould be carefully designed to reduce link establishmentdelay while considering energy efficiency. Two kinds ofhandshaking mechanisms are generally followed in Terahertzcommunication which are receiver-initiated and transmitterinitiated communication to establish links among differentnodes. In particular, the receiver-initiated MAC aimed atreducing the number of transmission in resource-constrainednano and macro-scale networks. Whereas the transmitter-initiated communication focuses on the performance efficiencyof the network in a traditional way. Typically, the directionalantennas are used for transmitter initiated communication inTCNs due to its narrow beam requirement and distance-dependent bandwidth [137]. Other than directional antennausage, some proposals use multiple antennas to establish initialcoordination between multiple nodes. The solutions so far onreceiver and transmitter initiated coordination protocols arementioned below and are shown in Figure 11.

A. Transmitter initiated MAC protocols

In a traditional way, the transmitter is mainly responsiblefor link establishment, data transmission and synchronization

of nodes parameters like scheduling times and channel infor-mation. Most of the Terahertz MAC protocols are followingthe transmitter-initiated communication due to its simplicityand distributed nature. However, the distance-dependent be-havior of Terahertz band due to absorption and path loss;directional antenna usage; high bandwidth and throughputsupport, increases the challenges. These challenges includethe antenna facing problem introduced when the transmitterand receiver remain unaware initially about the position andantenna direction; the hidden node problem; and reliabilityof communication i.e., the packet is lost due to path loss orcollision. A general example of transmitter initiated communi-cation is shown in Figure 12, in which a transmitter or a nodewhich has information can send a packet, a receiver whichreceives that packet can trigger an ACK or confirmation tosend and after that, a sender can send the Data packet. Thetransmitter and receiver can agree on the common parameterslike physical and MAC layer parameters to reduce the com-plexity of communication and to reduce the delay.

1) Nanoscale networks: In nanonetworks, mostly a nanocontroller is used to forward and collect data to/from nanode-vices in a centralized network. Transmitter-initiated commu-nication is used mostly to allow the nodes to transmit whenthey have data to send. The node which has data to send will

33

(a) Tx initiated communication

(b) Rx initiated communication

Fig. 12: Message transmission flow for handshake in TerahertzMAC protocols for link establishment. a) Transmitter initiatedhandshake mechanism which requires confirmation from re-ceiver of its sent packet before starting a data transmission,b) Receiver initiated handshake mechanism in which receiverinitiated communication when it required some information orhave enough energy to receive a message, mostly used in nanocommunication networks.

initiate the communication and perform handshake process.A transmitter initiated communication scheme is proposed

in [137] for nanonetworks, which is built on top of RD TS-OOK and takes benefits of a low weight channel codingscheme. The main aim is to negotiate between the transmitterand receiver the communication parameters and channel cod-ing scheme to minimize the interference and maximize theprobability of efficient decoding of the received information.The communication is established by sending a request by anode that carries information as Transmission Request (TR)and a node that receives it will agree to the communicationparameters and generate an ACK and sends a TransmissionConfirmation (TC) message. The TR contains the synchroniza-tion trailer, the transmission ID, packet ID, transmitting DataSymbol rate and Error Detecting Code. Although, it offersbenefits in terms of delay and throughput. It also has fewlimitations including the handshake process overhead whichlimits the Terahertz communication performance and limitedcomputational power of nanodevices which requires optimalcommunication parameters. Due to its limited computationalcapacity, in [139] energy and spectrum aware MAC protocol isproposed, in which the computational load is shifted towardsthe nano controllers. The works in which energy harvesting isalso considered and which uses Tx initiated communicationinclude [138], [154], [166]. Other works on Tx initiatedcommunication are mentioned in Figure 11.

The Transmitter initiated communication is used mainlyin papers in which distributed architecture is followed andwhere the goal is to achieve maximum throughput [151],[162] and not the energy efficiency. However, Tx initiatedcommunication can increase the control message overhead.As nanonetworks are limited in computational capacity, hand-shake mechanisms are required which can balance the energyharvesting and consumption with minimum delay for linkestablishment. The works which consider energy as wellare [138], [139], [154], [166]. A Tx initiated communicationis also followed in [166] which presents an energy model andalso considers the high node density.

2) Macro scale networks: For a macro-scale network withdirectional antennas, the steerable narrow beam is essentialto overcome the high path loss at the Terahertz band and toextend the communication range. Tx initiated communicationis used in many macro-scale applications, where main goalis to increases network throughput performance. A Terahertzbased communication model and autonomous relaying mech-anism for Vehicular communication are presented in [92], byconsidering the channel capacity, autonomous relaying andnetwork establishment. An antenna switching mechanism isproposed in [55], [90] for vehicular and small cell deploymentusing Terahertz band in which lower frequency bands are usedto achieve synchronization.

The approaches in [152], [158], [160], [168] uses multipleantennas to separate out the signaling and data transmissionsparts. The 2.4 GHz band is proposed to use for signalingand antenna alignment using an omnidirectional antenna toovercome the antenna facing problem. The initial access andcontrol information part are performed using IEEE 802.11RTS/CTS mechanism and for data transmission directionalantennas are used which also uses Tx initiated communicationto transmit data between the Terahertz nodes. Besides the highpath loss, the directional antennas in these works are shown tobe beneficial in reaching the distance beyond 1 meter. In [143],a pulse level beam switching is used with energy control butTerahertz band features are not considered.

A CSMA/CA-based channel access mechanism is usedin [84] with on-demand retransmissions mechanism for aTPAN which considers poor link conditions. The beaconframes are used by the piconet coordinator to provide infor-mation like channel access, channels slot assignment, and syn-chronization information. It is shown that network throughputdecreases when the channel conditions are poor and proposedMAC protocol shows better performance in comparison withIEEE 802.15.3c and ES-MAC [139].

Although, the transmitter-initiated protocols are widely usedas they incur less complexity and favors the distributed nature.It incurs several challenges due to the use of the directionalantenna. On one hand, these directional antennas increasethe transmission distance. On the other hand, they introducethe antenna facing the problem. In outdoor scenarios wheremobility is involved frequent beam switching occurs whichrequires a novel mechanism to minimize the synchronizationand antenna alignment schemes. The WiFi technology isproposed to minimize the control messages overhead and linkestablishment in works like [152], [158], [160], [168] and

34

for antenna alignment which overcomes the facing problem.However, it introduces the high antenna switching overheadand requires efficient scheduling mechanism for seamlesscontrol information and data dissemination transmission.

B. Receiver initiated MAC protocols

A general example of the receiver-initiated communicationis shown in Figure 12, in which the receiver announces itsexistence and readiness to receive a packet from the sender.The Rx initiated communication is mainly used in nanoand macro scale networks to save energy and reduce excessmessage overhead. Different existing solutions for both typesof networks are discussed below.

1) Nanoscale networks: In nanoscale networks, whichare prone to energy utilization, the excess of transmissionsor message exchange means more utilization of energy. Innanonetworks, the amount of energy stored is just enoughto hardly transmit a packet [144], [237]. The transmissionremains unsuccessful when the receiver receives the requestfrom the sender but does not acquire enough energy to send anACK or data packet. Therefore, in receiver-initiated protocols,the receiver takes initiative and announce its status of energyfirst to all senders by sending a Request-to-Receive (RTR)packet. The energy efficiency and harvesting are also discussedin Section IV-B2.

The receiver initiated communication is used both in cen-tralized as well as distributed networks [127], [144], [146],[147], [244]. In centralized nanonetworks, a nano controlleris mainly responsible for performing major processing anddecision making, as they are energy enrich devices. Since thereceivers are usually assumed to generate their own energyresources, they harvest just enough energy to transmit apacket. Therefore, in solutions like [127], [144], [146], thereceiver starts the communication by informing the senders,when it is ready to receive a packet and can exchange theinformation like coding schemes, error rates, and scheduling.Therefore, limited energy resource is one of the main reason,receiver-initiated communications are preferred in centralizednetworks. The problems occur when the receiver remains busyenergy harvesting phase, and senders start sending the packets,which can result in the loss of a packet. Therefore, schedulingbecomes an essential part of such kind of schemes. In [146],a receiver-initiated communication model is presented forcentralized topology, in which a receiver announces an RTRpacket to nearby nano nodes and then the nano nodes send aDATA packet or ACK in response in a random-access mannerwith probability p to establish a handshake between the nodes.A distributed scheme is presented in [127] in which schedulingand harvesting mechanisms are proposed to work together toenhance the energy utilization of nanonetworks. The proposedscheme uses the receiver-initiated approach to achieve thehandshake and schedules.

The unsuccessful transmission-reception because of thenode harvesting phase can increase dalay and also cause ahidden node problem. Therefore, new schemes are requiredto avoid the hidden node problem in a distributed environ-ment using a receiver-initiated approach. A MAC protocol is

discussed in [144] in which optimal energy consumption andallocation problem are presented which aims to maximize thedata rate. It is also shown that the amount of energy harvestedis not enough for transmitting one packet and therefore it cantake multiple timeslots to transmit a single packet when theharvesting rate is lower than the energy consumption rate.

In nanonetworks, receiver-initiated communication providesthe flexibility to nodes in deciding which when to receive andtransmit. The message overhead in these networks can be re-duced by following a one-way handshake as provided in [147],[244] in which energy harvesting is used. Whereas in [144]the trade-off between energy harvesting and consumption isalso discussed.

2) Macro scale networks: In these networks, high pathloss at Terahertz frequencies affects the achievable distancebetween the Terahertz devices (cf. Section IV-A). It alsorequires tight synchronization between a transmitter and re-ceiver to overcome the deafness problem [147], [247]. Areceiver-initiated MAC protocol using directional antennas isdiscussed in [147], [244] which uses a sliding window flowcontrol mechanism with a one-way handshake that increasesthe channel utilization. High speed turning directional antennasare used to periodically sweep the space. The main objectiveis to prevent unnecessary transmission when the receiver is notavailable due to antenna facing the problem. In this scheme,a node with sufficient resources broadcasts its current statusby using a CTS message by using a dynamically turningnarrow beam while sweeping its entire surrounding space.The CTS frame contains the information of receivers’ slidingwindow size. On the other side, the transmitter checks fora CTS frame from the intended receiver and then pointsits direction for the required period towards the receiver.The initial neighbor discovery of the neighbor nodes is notconsidered in [147]. Further, due to the bit error rate and pathlosses (cf. Section IV-B1), packet reception guarantee is notconsidered. It is also possible that multiple transmitters mightpoint out at same receiver at the same time, which can resultin possible collisions. A MAC protocol focused on cross-layer analysis for relaying strategies is discussed in [163] withdistance dependant throughput maximization analysis whileconsidering the antenna, physical, link, and network layereffects.

In particular, receiver-initiated communication is shown tobe better than the transmitter-initiated communication in [147],[244]. In general, the receiver-initiated communication ismostly used in nanonetworks due to its limited resources.However, it is also been proposed in Macroscale networks toreduce the message overhead and to solve an antenna facingproblem. In distributed environments, pre-assignment of rolescan increase the delay in terms of achieving coordinationand can also cause the hidden node problem. Although,Rx initiated schemes can reduce the message overhead thecomplexity increases with assigning extra responsibilities fora receiver.

C. Summary and discussionThe nanonetworks are considered as networks with limited

energy and resources in which mainly the schemes are pre-

35

ferred in which energy consumption can be minimized. Otherthan the traditional way, some receiver-initiated mechanismare proposed in the literature, in which a receiver initiates thecommunication establishment by announcing its status for thesufficient energy resources to receive a packet. Whereas, intransmitter initiated communication, the node which have in-formation to send can initiate to establish the communication.At nanoscale networks, mostly the networks are centralizedand therefore nano controllers are used mostly to manage thecontrol and data transmissions and so form a centralized net-work. In this kind of network, scheduling the transmission isthe main requirement as they use pulse based communicationinstead of carrier sensing based communication. Although,the bandwidth is high and the probability of collision is low.But, that can be true only when the network density is low.For higher network density, the collisions cannot be ignoredand so require an efficient mechanism to avoid collisions andminimize packet loss.

The receiver initiated communication is used also in themacro scale networks. Although it minimizes the controlmessages overhead, it increases the complexity and is notpreferable in the distributed scenarios. In a distributed environ-ment, where two nearby nodes performing the same operationsuch as energy harvesting can increase the delay and can causea hidden node problem. Therefore, efficient synchronizationand scheduling mechanism are required for these scenarios.

The Terahertz band in itself is sensitive to atmosphericmolecules and so the path loss is high and it increases morewith an increase in the distance. Efficient mechanisms arerequired to achieve the reliability of transmissions, whichcould distinguish between packet error, collisions and packetloss due to other reasons including medium uncertainty. Sofar, the single-channel mechanism is preferred in the Terahertzcommunication network. However, flexible MAC mechanismis also required which can work in multiple channels scenarioas well as shared access in both single and multiple channels.

VIII. CHALLENGES AND FUTURE RESEARCH DIRECTIONS

Designing an efficient Terahertz MAC protocol needs toaddress different challenges. In this section, these challengesare presented with future research directions.

A. Terahertz communication network topologies

1) Macro scale network: The challenges and future re-search directions for macro scale network topology design arediscussed below.

Static and mobile scenarios: In fixed point-to-point con-nectivity, it is fairly easy to maintain stable links. However,when the nodes are mobile, the network topology changesfrequently, in which frequent neighbor discovery and linkestablishment will be required. To move further, point-to-multipoint connectivity requires more attention in terms ofMAC protocols design. A Data Centre environment is a goodexample, in which establishing point-to-point and multipointamong the inter and intra rack communication to effectivelyreplace wired connections is still a challenge.

The Terahertz signal is highly sensitive to the outdoor en-vironment, for example, weather fluctuation, and the presenceof blockers between two nodes can affect the communicationlink. MAC layer for such a system should include fast linkre-establishment mechanism and alignment operation in caseof sudden miss-alignment between the two nodes by givingalternative paths for transmission.

Blockage and beam steering: Antenna arrays and reflectorscan be used to avoid the chances of blockage and to find a goodpropagation path. It is critical to find a good beam pointingmechanism to support user mobility with fast and accuratebeam tracking, both in LOS and non-LOS conditions [248].A mechanically steerable antenna is demonstrated in for 300GHz band [249]. Overall, a faster electronic beam steering ismore practical and required.

Unlike lower frequency band such as microwave, Terahertzis very sensitive to blockers. Efficient techniques are requiredto reduce the effects of blocking in the system. The pathdiversity can be one of possible solution using beam diversityand MIMO system. The challenge is to track presence ofblockers and to keep the link available when direct LOS isabsent. The integration of Terahertz intelligent surface is worthconsidering which enables user tracking and add path diversityto the propagation scenario, users can be reached even inpresence of severe blocking scenario.

Data Centre Geometry: The high capacity Terahertz linkscan help in re-designing the Data Centre geometry by movingthe racks closer in a cylindrical architecture. Future workdemands the flexible Data Centre architecture to supportscalability and energy-efficient design while considering theTerahertz features and limitations. It will reduce the deploy-ment delay and cost for future data centers. Full beam scanningcan be used for initial access to serve 360-degree search space.Antenna sectors can also be used for 90-degree search spaceand communication at the cost of additional hardware.

Coverage: The transmission distance achieved so far forTerahertz communication is still limited to few meters [22].Extending this range to reach to 10 m or more for an indoorscenario using low power transmission devices and efficientcommunication protocols is an active field of research forB5G networks. As a future topic, the Terahertz wirelessand fiber connectivity for backhaul and fronthaul high datarate communication are also gaining attention (cf. Sect.III).Directional antennas with narrow beams are being encouragedin Terahertz networks to extend the transmission distance. Thisdirectional antenna usage can limit the interference and losses(cf. Section IV-A), they must use the optimal schedules andtight synchronization. The phased array and MIMO techniquescan be used to further extend the transmission distance.

Terahertz communication is characterized by a low coveragezone as transmitted signal experiences high path attenuation. Ahigh gain antenna can be deployed to increase network range,however the communication range is still low compared tolower frequency bands. In order to extend coverage zone itis possible to add additional functionalities to nodes at thecoverage area zone to play the role of coverage extension andcoordination with nodes out of the coverage zone. Coverageextension is challenging as more interferences occur and

36

Fig. 13: Coverage extension in Terahertz networks.

additional time is required for out of range nodes discoveryis required, some related work on coverage extension can befound in [22]. From MAC layer point of view, nodes at theedge should coordinate to discover and synchronize with out-of-range nodes. A second possibility to further extend thecoverage area at macro scale is the deployment of intelligentsurfaces [250]. In which new device component that canreplace traditional reflectors and lenses, whenever a node inthe inner coverage zone is near the intelligent reflector. It cansend discovery messages and data to the surface in order tobe reflected out of the inner zone. The MAC layer shouldbe aware of new devices required for coverage extension.Edge nodes are selected by the nodes controller for coverageextension if their received power is lower than a thresholdand also if they are located near an intelligent surface. Thisconcept is shown in Figure 13.

2) Nanoscale network: The challenges and future researchdirections for nanoscale network topology design are discussedbelow.

Node density: In the nano-communication network, thou-sands of nodes can exist in very small regions which aregenerally limited in capabilities. These networks require ef-ficient MAC layer techniques to meet network requirementssuch as high throughput, reliability, and low energy consump-tion. Models related to node density are used in literature,for example interferers are linearly distributed around thereceiver [251]. The density distribution is used to evaluatethe interference in the system and assess the limitation of thenetwork, non-homogeneous density of nodes can affect theMAC choice, then MAC protocols should be able to track thefast change in node density in each part of the network. Afurther assumption on network density can be used such asthe point Poison process model.

With limited energy storage capacity, establishing linksand data transmission will require more energy to facilitatelarger number of nodes. The nodes mobility in a distributedenvironment can also change the topology of the nodes fre-quently which needs to be handled using new MAC layerprotocols. In different applications, like agricultural or in-body health monitoring, the nodes scalability is required(cf.Sect.III-B). Managing node discovery and link stability arestill challenges while dealing with limited energy resources atnanoscale. Models to quantify the resources like time, storageand amount of communications are required to reduce thecomputational complexity in centralized as well as distributed

networks. Managing scalability in distributed networks is alsoa challenge.

Energy harvesting and transmission trade-off: The limitedcapacity can allow only an energy harvesting or transmissionat one time which requires to be scheduled efficiently. Whena node busy in harvesting it can not transmit and when ittransmits it can not generate energy for next transmissionswhich can affect the overall latency and throughput of ananonetwork. Therefore, efficient schedules are required toimprove the overall throughput and latency while managingthe harvesting and transmissions.

Computational complexity: Nano devices are small deviceswith limited computational capacity and power managementand so require simple communication protocols. More func-tionalities can be provided by optimizing the nano device’sbehavior. Therefore, to manage the nanodevices and process-ing the complex computations externally, hierarchical archi-tecture is required with decisions to be taken at devices withhigher computational capacity and powerful communicationlayers [252]. Gathering the data in a distributed networkarchitecture to perform complex operations is a challengingtask. Therefore, new mechanisms are required for synchro-nization and coordination for distributed networks. Further,the nanonetworks can also be combined with current networktechnologies like software-defined networks, IoTs and virtualnetworks to solve the complex problems at those technologiesand higher layers [18].

High capacity controller: The nanonetwork topology canbe dynamic at times, due to dynamic channel conditionswhich can affect the reliable transmissions. In response, thenanodevices in a network might not share the same topologyinformation. Further, due to limited memory, storage, and com-putational processing capabilities. The nanodevice also facesdifficulty in storing the routing tables and heavy algorithms.To solve this, a controller with software-defined functionalitycan be utilized to do the extra computations, which can alsochange and reconfigure the behavior of nanonetwork.

B. Terahertz channel access mechanisms1) Macro scale network: Since traditional channel access

techniques can not be directly used in Terahertz networks dueto unique features of the Terahertz band (cf. Section IV-A).New and novel techniques are required to provide highdata rate and reduce the transmission delay. In this section,challenges and future research directions are mentioned forefficient Terahertz channel access mechanisms.

Hybrid protocols: The scheduled channel access techniquescan reduce the collisions, whereas, random access techniquescan enhance the throughput and latency requirements. Thefuture research should be on hybrid mechanisms to improvethe fairness and throughput among the users using combinationof both techniques.

Adaptive band selection: The Terahertz transmission bandscan be affected by high absorption and path loss. Multipletransmission windows are available in the Terahertz band withdifferent attenuation values. Efficient mechanisms are requiredto efficiently utilize the distance and bandwidth dependentTerahertz band allocation.

37

Multiple band usage: To solve the synchronization andinitial access, different bands like microwave, mmWave, andTerahertz are being used together. The microwave band canprovide wider coverage which can be used for initial accessand antenna alignment. However, when to switch among thesebands requires new methods to be defined to improve theswitching overhead and delay.

Multiple antennas and beam usage: MIMO antennas canbe used to mitigate the Terahertz channel effects (cf. Sec-tion IV-A) and can improve the path diversity, and antennagain which can improve the overall coverage.

High order modulation: New waveforms, and modulationand coding schemes for the Terahertz system are also requiredto improve the data rate and quality of service for beyond 5Gcommunications [253]. The use of OOK mapped with ultra-short pulses at one hand can reduce the transceiver complexity,however, it also introduces challenges for antenna arrays forultra-broadband design. The complexity of generating andprocessing high order modulated signals at transmitter andreceiver side using Terahertz device is a challenge. For goodchannel conditions, throughput can be improved by using thesehigh order modulation techniques like 64 QAM. Further, thedecoding and delay complexity at receiver also needs to befocused.

Beam management: When using particular antenna archi-tecture with directional beams, then how to map betweenaccess time, frequency and beam direction. Fast beam scanningand steering mechanisms are required to obtain channel stateinformation to improve the channel usage decisions at macro-scale networks. The random channel access can solve the prob-lem of antenna alignment, however, the range becomes shortdue to high path loss. An efficient TDMA based schedulingmechanism is required for both centralized and distributednetworks to avoid multi-user interference.

2) Nanoscale network: In nanonetworks, due to shorterdistance and low path loss huge bandwidth can be used whichcan result in very high data rate, short transmission timewith low collision probability [18]. The MAC protocol roleis very important in regulating the link behavior, arrangingthe channel access and coordinating the transmission in adistributed environment.

Random access and scheduling: At the nanoscale, hugebandwidth availability and pulse-based communication, re-duces the collision probability. The random access techniquescan take benefit out of it but require efficient scheduling oftransmission with a larger duration between packet transmis-sions. However, for high node density scenario, this needs tobe carefully designed due to energy limitations [89].

Node density and error control: Although the actualTerahertz nanoscale network is characterized by a basic MAClayer, it is possible to implement efficient error control pro-tocols to reduce packet retransmissions and packet waitingtime. Optimized mechanisms are required overall to run com-munication tasks at the nanoscale, due to energy limitationsand to support a large number of nanodevices. Efficientscheduling mechanisms for nanonetworks are required overallto balance the energy harvesting and channel access for datatransmissions. They need to be further optimized, to enable

fast communication among a larger number of nanodevices.Further, additional functionalities such as scheduling trans-missions are required between inter-cluster and TDMA basedintracluster communication to avoid collisions and to increasethroughput [254].

C. Terahertz Receiver and Transmitter initiated communica-tion

1) Macro scale network: Message overhead: Mainly,the transmission of control messages in excess can causethe message overhead problem which can easily occur indirectional Terahertz communication. Efficient techniques arerequired to reduce the message overhead. Different windowsor bands can be used to utilize the channel resource efficiently.Although energy is not a high constraint like in nanonetworks,fast link establishment techniques, especially for distributedTerahertz networks and networks with high mobility requirenovel solutions, for example, small cells and vehicular com-munications. The control packets transmission to establisha link can cause high message overhead. New handshakingmechanism is required to establish link with reduced messageoverhead and delay.

Memory assisted communication: A memory assistedcommunication can enhance the throughput and latency ina Terahertz network [158]. Nodes position, neighbor tables,beam directions, and weights can be used in memory assistedcommunication to reduce the alignment and training time forantenna. Relays can be selected by coordinating the neighborinformation and transmissions can be scheduled in advancewhile tracking the node’s availability.

Antenna direction and communication: At macro scale,communication is linked with directional antennas. To initiatecommunication, nodes should have enhanced MAC functional-ities such as simultaneous sensing and beam steering capabili-ties to track channel and node position jointly. Additional tech-nology can assist in transmission initiation, such as RADARand LIDAR for mobile nodes, and lower band technologies.The radar-based sensing is gaining attention to solve thedeafness problem in the vehicle to vehicle communicationto efficiently align the antenna direction and communicationestablishment [94]. A preamble injection is shown useful inlow SNR scenarios to avoid deafness problem in [93].

2) Nanoscale network: For nanoscale networks, nodes withhigher energy harvesting capabilities can take the responsibil-ity of link establishment and transmission initiation, howeverdesigning sophisticated algorithms for rapid link initiationstill remains an issue. Research is required for distributedMAC protocols for Nanosensor networks to reduce messageoverhead with energy control mechanism. Receiver initiatedtransmission suits the low power nanodevices link establish-ment process due to reduced message overhead, it shouldconsider energy as well as data size to be transferred. Acombination of distributed communication in presence of acontroller can help in managing the transmission schedulesamong nano devices [255].

38

D. General challenges and future research directionsThe Terahertz technology is capable of delivering high data

rate traffic with an acceptable quality of services such as lowdelay, and minimized energy consumption. However, manychallenges still exist and require further research and attention.Some of them are discussed below related to Terahertz MACprotocols in general.

1) Interference management: Although interference canbe reduced by using highly directional antenna, it is notconsidered by many research works [35]. It can be consideredfor large network requiring high data rate connection pernodes, such as top of the rack data center network, wherenodes should transmit all the time with high data rate and lowdelay. Interference for dense indoor scenarios should be deeplystudied and interference model needs to be established. TheMAC layer should be aware of interference in the channel toelaborate further the rapid scheduling and fast channel accessand switching based on channel interference information. Thenew and dynamic channel selection mechanism is also requiredwhile considering the Terahertz band’s unique characteristicsand band-specific interference and achievable distance.

The interference management module can track channelstatus and decides on the transmission time slot as well asthe carrier to be used and physical parameters to be set,such as modulation and coding scheme and annulating sidelobes in some directions. The additional procedure can be alsoimplemented such as adoption, at the design stage, of a specificfrequency plan for network and setting a sub-band spacingstrategy for each application. At operational mode, each nodecan use a fixed frequency pattern at the deployment stage oradopts the frequency hopping strategy to keep an acceptableSINR and to overcome the molecular absorption of the prop-agation medium [145], as the noise generated by moleculedepends on frequency. Using frequency hopping scheme ispromising as it tracks the channel switching, however, thedesigning of the frequency hopping algorithm is a challenge.The second challenge is to explore the number of frequenciesthe MAC layer can manage to improve the throughput.

2) Antenna design: The link quality depends on the phys-ical layer and channel, for Terahertz communication antennatechnology improvement is considered as a key factor for linkbudget enhancement. The MAC layer should monitor antennaby fast switching beams to serve all users in a fair way. MACcan also select antenna carrier frequency and polarization. Tomonitor efficiently a Terahertz communication, MAC layershould interface with the antenna system, for example con-trolling the steering angles, beam width and switching time.Because a good command of antenna system can increasedata throughput and reduce delays due to misalignment errors.Antennas properties should be optimized, the MAC layer canmonitor Terahertz antenna via frequency switching, beam-forming, and diversity in order to meet network requirementsincluding:

Polarization capabilities: Using two polarizations (hori-zontal and vertical) with sophisticated algorithms of cross-polarization cancellation is promising to boost the Terahertzsystem performance toward higher throughput and lower totalsystem signal to interference ratio [256]. The main challenges

with the dual-polarization approach from the MAC point ofview are to balance the traffic between the two polarizationsand to mitigate errors. Moreover, channel impulse responseand then received power depends on polarization [257], onechallenge is how to exploit efficiently Terahertz wave polar-ization to increase data throughput by balancing the data flowsimultaneously between horizontal and vertical polarization.

Wideband and multi-band antenna: The design of multiwideband antennas are required to increase MAC efficiencyand meet system requirements in term of high throughput,as more bandwidth will be available to transmit more datarate [258]. Using a multiband antenna can also reduce systemlatency by deploying separate bands for data transmission andcontrol message exchange.

High antenna gain: To mitigate channel impairment andextend the communication range, antenna gain should bemaximized, Horn, logarithmic antenna and phased array arepromising for designing high antenna gain Terahertz commu-nication. For high antenna gain, it is possible to increase nodereachability, but more care should be addressed to antenna sidelobes as they can generate more interferences.

Spatial diversity: To mitigate channel impairment andincrease channel capacity, multiple antennas along with phasedarray, exploiting Terahertz propagation diversity, can be de-ployed for Terahertz links, such as MIMO and ultra massiveMIMO. Using MIMO increase spectral and energy efficiencyfor the link, however, it requires efficient signal processingperformance to encode and decode MIMO signals and exploitdiversity. From MAC point of view, deployment of ultra mas-sive MIMO will affect resource allocation techniques [210].

Fast switching capability: To increase data throughput andreduce latency per link, the beam switching time should beminimized. Switching can occur at pulse, symbol or framelevel.

Adaptive beamforming: Directional antenna is consideredan alternative solution to mitigate channel impairment andincrease link budget, nevertheless, antenna pattern can takedifferent shapes, and it should be optimized for a networkuse case, by using adaptive weighting of antenna elementsmonitored at the physical layer and MAC, it is possibleto reduce effect of interference by annulling lobes in somedirection to avoid interferences. A MAC module can beimplemented to control antenna beamforming and adapt theantenna technology to the network topology. In [259], a log-periodic toothed antenna is optimized for beamforming andbeam steering for the Terahertz band. A concept of intelligentcommunication by using ultra massive MIMO is proposedin [260], to increase the communication distance and data ratesat Terahertz band frequencies.

3) Synchronization: Synchronization adds accuracy to net-work operations and coordination and reduces frame collisionsamong nodes, as a result, it contributes to QoS enhancement.Moreover, it is responsible for more computation complexityand requires additional time slots before data transmissionstarts. Nodes memory and time for link establishment are themain cost to pay, in order to deploy the synchronous network.At the MAC level, the challenge is to design algorithms fornodes and frame synchronization, for nodes to be aware of the

39

transmission time. Another challenge is to efficiently allocatethe radio resources for synchronization procedures such asfrequency, time and power, which can increase the delay. Toreduce the delay transceivers with more capabilities such asmemory and processing can be used.

4) Transceiver design: With sophisticated antenna Tera-hertz design and efficient physical layer functionalities, achallenge related to transceiver performances should be con-sidered. MAC layer aims at adapting the high data rate trafficsto the physical layer. An efficient transceiver is required to dealwith different MAC functionalities ranging from framing, syn-chronization, error control, scheduling and buffering. Authorsin [159], demonstrate that it is possible to optimize transceiverarchitecture to bear high data rate reaching 100Gbps usingparallelism and optimized memory usage along with framelength and error control techniques. Using efficient processingtechnique at the transceiver and sufficient memory size, itis possible to implement MAC protocols dealing with fastchannel access, efficient scheduling technique and multi trafficcommunication

5) Link establishment, neighbor discovery and deafnessproblem: Before any communication starts, an establishmentphase should be initiated. Link establishment is the duty of theMAC layer when a node needs to transmit to another node.This phase starts with setting up all the required parameterssuch as physical layer parameters, timers, synchronizationprocedure, then after receiving the acknowledgment from thereceiver, a new transmission can begin. The challenge is howfast we can establish any Terahertz connection, and how toincrease the success probability of link establishment phase.

Deafness can complicate the neighbor discovery due totransceiver misalignment and prevents the control messagesto be exchanged in a timely manner (cf. Sect.IV-B2). Toavoid the antenna miss alignment and link establishment, themmWave standard IEEE 802.15.3c and IEEE 802.11ad usebeam training, in which one node operates in a directionalmode and other node search space in a sequential directionalmode. After a complete search, sector-level training occursto perfectly align the beams. For neighbor discovery thechallenges are to discover all the nodes with minimum delay,and techniques to search the space for beam alignment in ashort time when two nodes are not initially aware of theirbeam directions.

Particularly, for Terahertz communication the neighbor dis-covery is challenging due to unique band characteristics andantenna directionality, and for nanoscale networks due tolimited energy resources. Neighbor discovery is required tosynchronize nodes within a network and rapidly considernew nodes in the network. As a future research direction,optimization of discovery time by correctly choosing referencetime for antenna alignment is an important challenge withtimely information exchange. The discovery can be enhancedby using multibeam with fast switching and coordinationamong distributed nodes [261]. A neighbor discovery protocolwith directional antennas with side lobes information andfull antenna radiation pattern to better detection is proposedin [244]. However, multipath effects and LOS blockage werenot considered.

6) LOS blockage: A blockage is a situation when an objectcrosses the main link between two nodes transmitting to eachother, it can also be generated from frequency shifting andreflected signal from surrounding objects. Due to high datarate, a small or temporary blockage can result in very highdata loss. Therefore, it is important to propose novel anti-blockage mechanisms to avoid blockage situations and toachieve seamless coverage. At MAC layer, it is importantto identify blocked channels to avoid false detection andcorrection and to distinguish between deafness and unblockederror. In Terahertz band, due to small wavelength (0.3 mm) thedirectional links can be easily attenuated by LOS obstacles. Inmobility scenarios, these obstacles can occur more frequently,and therefore can degrade the Terahertz link performance.Only increasing the transmission power cannot penetratethe obstacles, therefore an alternative unblocked channel isrequired to steer around. Reflectors can be used to avoidpermanent blockage, but new mechanisms are required withbeam steering and management functionalities to avoid linkblockage. Modeling the blockage phenomena for each usecase is required, and MAC layer should be aware of it. Themain challenge is to detect blockage and tackle this issueat MAC layer. One alternative is to differ its transmissiontill the link is cleared or to select an alternative path withnew parameters to avoid transmission interruption. To mitigatetemporary blockage of LOS wireless link, an NLOS wirelesslink with reflection and scattering over multiple rough surfacesis analyzed in [262]. The use of the NLOS links can broadenthe access point options to improve the link performance.However, using complex modulation schemes are yet to beanalyzed for its feasible working under different environments.

7) Design of relaying and multihop protocols: The short-range communication like indoor and Data Centre scenarios,require new and efficient relaying techniques to increase thereachability of nodes. Nodes relaying or forwarding capabilitycan be implemented at the MAC layer. It is activated whenthe signal from one transmitter needs to be regenerated byintermediary nodes to reach its final destination. However, toactivate this, each node must have a complete view of theneighbors which can be exchanged among the nodes as aneighbor table. Due to antenna directionality and tight beamrequirement of Terahertz communication, beam switchingtechniques can be used where antenna can take 0.1 ns to switchantenna beam direction [212], which can increase the overalldelay in forwarding the packets. The relaying protocol usingdirectional antenna must be designed to reduce this delay andto overcome channel impairment problem.

In Terahertz band only short transmission range is achiev-able until now. Due to which the signal needs to regeneratedby an intermediate node to reach the destination. Designingstrategies with relaying capabilities by considering the uniqueband features and environment is a challenging requirement.Work on node relying on Terahertz band was performedat nanoscale communication [263], where two modes wereconsidered: amplify and forward, and decode and forward, tostrengthen the direct path by maximum ratio combining.

The relay node can be selected from the existing neighborsor can be placed especially in a network. Each mechanism

40

needs to address different challenges including link quality,location, and number of relays. In a distributed environment,where nodes communication range is shorter, multihop com-munication should be enabled. Multihop protocol design canbe challenging when high throughput and low latency are therequirements. Therefore, new multihop strategies are requiredto fulfill Terahertz communication requirements by consider-ing limited capabilities and behavior of communication layersin case of nanoscale networks. For macro-scale networks, thepath loss, molecular noise, and antenna directionality shouldbe considered. Reflectors can also be used for communicatingand reaching to nodes with LOS blockage.

8) Coordination: Designing efficient MAC protocols forvehicular and satellite communication with relaying and coor-dination among the nodes is an open challenge for MAC de-sign. The node should decide the next relay node to strengthenthe communication among different nodes using coordinationmechanisms. The nodes should be capable of deciding whichnode it should coordinate.

9) Cross layer design: Terahertz cross-layer design forservice requirements can be performed using physical andnetwork layer aware algorithms at the MAC layer. MAC layershould adapt between traffic flows coming from the upperlayer and the channel fluctuations along with the physicallayer procedures. For instance, the selection of frame sizeper transmission period should be adaptively chosen based onpacket arrival from the networking layer as well as consideringmeasurements from the physical layer. Scheduling transmis-sions can be optimized based on measurements gathered fromthe physical layer. MAC decisions such as band selection andpath switching, in case of blockage are also affected by thephysical layer and channel status. The transceiver memoryshould be optimized to support traffics with different QoSprofiles.

10) Scheduling: In the Terahertz network, scheduling algo-rithms can enhance the overall quality of service by using radioresources for a given policy such as maximizing throughput,minimizing the total interference in the network or reduc-ing system delay. The scheduling module will be interfacedwith the medium access module as well as physical layer,where knowledge about the channel conditions and trafficrequirements will govern scheduler decisions. Exchanginginformation related to buffer, channel quality and requirementof each traffic flow should be considered by the scheduler andalso schedules of other nodes.

11) Framing and Error handling: Selection of frame size,frames, and multi-frame structure and error control strategies,such as CRC insertion and frame retransmission, can enhanceTerahertz link in terms of frame error rate as well as leadsto an increase of throughput. Adaptive frame size and controloverhead are fundamental to maintain a communication linkand reduce errors among transmitted frames. Increasing thepacket size can cause a higher number of channel errors whichrequires more robust error detection and correction schemes.The longer packets can also introduce the buffering problems.Therefore, the optimal packet size and analysis of the trade-offbetween the size and performance, and flow control policiesto avoid congestions and buffer overflow, requires further

research.12) Mobility management: Mobility can easily affect the

quality of the established links due to narrow beams. There-fore, frequent re-establishment of links is required to maintainthe links and communication over it. Two mobility models arementioned in [264], linear and circular motion. In differentTerahertz scenarios, different mobility models need to bestudied like in V2X scenarios and small cells, where LOSblockage can also occur frequently. It is important to trackthe best beam in case of frequent link breakage and beamalignment requirement.

In V2X networks nodes change their position with variablespeed. To keep the connectivity with the network, a man-agement module should be implemented at the MAC layermonitoring node location and tracking its speed. For suchan application, a mobility model needs to be proposed foreach scenario, and how MAC layer should keep trackingnodes position as well as its neighbors then how to selectand reselect nodes to which it should transmit while takinginto consideration also the receiver mobility. The mobilitymanagement module will decide on the handover and howto make the link robust all the time until the end of thetransmission without interruption. It is possible to decide tochange a new node as receiver or to transmit by relaying. Anupdate mechanism should be set to sort all neighbor nodesbased on their availability. One more challenge is how tosample channel conditions and how fast the MAC can decidefor the handover.

IX. CONCLUSION

In this paper, a comprehensive survey is presented forTerahertz MAC protocols. The Terahertz band has a greatpotential to support the future ultra-high bandwidth and lowlatency requirement for beyond 5G networks. Especially, theexisting unused frequency operating at disruptive bandwidthsof 70 GHz can be a key enabler for beyond 5G networks. Inthis regard, the key features, design issues for Terahertz MACand decisions which should be taken at MAC layer to enhancethe performance of the network, are highlighted and discussed.Different Terahertz applications for macro and nanoscale arealso discussed with their scenario-specific challenges. Thesurvey has identified numerous gaps and limitations of existingTerahertz MAC protocols for enhancing further research in thisdomain. To highlight the limitations, the existing literature onTerahertz MAC protocols is also classified based on topol-ogy, scale, channel access mechanism and transmitter/initiatedcommunication. To push further the research in this domain,challenges and future research directions are also presentedwith a cross-layer approach.

ACKNOWLEDGMENT

This project has received funding from the EuropeanUnion’s Horizon 2020 research and innovation programmeunder grant agreement No 761579 (TERAPOD).

41

REFERENCES

[1] C. Systems., “Cisco visual networking index: Globalmobile data traffic forecast update, 2016âAS2021white paper,” Tech. Rep., 2018. [Online]. Avail-able: https://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/mobile-white-paper-c11-520862.html

[2] A. S. Hamza, J. S. Deogun, and D. R. Alexander, “Wireless commu-nication in data centers: A survey,” IEEE Communications Surveys &Tutorials, vol. 18, no. 3, pp. 1572–1595, thirdquarter 2016.

[3] I. F. Akyildiz, J. M. Jornet, and C. Han, “Terahertz band: Next frontierfor wireless communications,” Physical Communication, vol. 12, pp. 16– 32, 2014.

[4] ——, “Teranets: Ultra-broadband communication networks in the Tera-hertz band,” IEEE Wireless Communications, vol. 21, no. 4, pp. 130–135,August 2014.

[5] H. Elayan, O. Amin, R. M. Shubair, and M. S. Alouini, “Terahertzcommunication: The opportunities of wireless technology beyond 5G,”in International Conference on Advanced Communication Technologiesand Networking (CommNet), April 2018, pp. 1–5.

[6] T. Yilmaz and Ö. B. Akan, “On the 5G wireless communications at thelow Terahertz band,” CoRR, vol. abs/1605.02606, 2016.

[7] H. Zhang, S. Venkateswaran, and U. Madhow, “Channel modeling andMIMO capacity for outdoor millimeter wave links,” in 2010 IEEEWireless Communication and Networking Conference, April 2010, pp.1–6.

[8] S. K. Yong and C.-C. Chong, “An overview of multigigabit wirelessthrough millimeter wave technology: Potentials and technical challenges,”EURASIP Journal on Wireless Communications and Networking, vol.2007, no. 1, p. 078907, Dec 2006. [Online]. Available: https://doi.org/10.1155/2007/78907

[9] T. Kurner and S. Priebe, “Towards Terahertz communications status inresearch, standardization and regulation,” Journal of Infrared, Millimeter,and Terahertz Waves, vol. 35, no. 1, pp. 53–62, January 2014.

[10] I. F. Akyildiz, J. M. Jornet, and M. Pierobon, “Nanonetworks: A newfrontier in communications,” Communications of the ACM, vol. 54, no. 11,pp. 84–89, 2011.

[11] ——, “Propagation models for nanocommunication networks,” in Pro-ceedings of the Fourth European Conference on Antennas and Propaga-tion, April 2010, pp. 1–5.

[12] F. Dressler and F. Kargl, “Towards security in nano-communication:Challenges and opportunities,” Nano Communication Networks, vol. 3,no. 3, pp. 151 – 160, 2012.

[13] J. M. Jornet and I. F. Akyildiz, “The internet of multimedia nano-things,”Nano Communication Networks, vol. 3, no. 4, pp. 242 – 251, 2012.

[14] I. F. Akyildiz and J. M. Jornet, “The internet of nano-things,” IEEEWireless Communications, vol. 17, no. 6, pp. 58–63, December 2010.

[15] K. Darchini and A. S. Alfa, “Molecular communication via microtubulesand physical contact in nanonetworks: A survey,” Nano CommunicationNetworks, vol. 4, no. 2, pp. 73 – 85, 2013.

[16] D. Malak and O. B. Akan, “Molecular communication nanonetworksinside human body,” Nano Communication Networks, vol. 3, no. 1, pp.19 – 35, 2012.

[17] I. F. Akyildiz and J. M. Jornet, “Electromagnetic wireless nanosensornetworks,” Nano Communication Networks, vol. 1, no. 1, pp. 3 – 19,2010.

[18] Q. H. Abbasi, K. Yang, N. Chopra, J. M. Jornet, N. A. Abuali,K. A. Qaraqe, and A. Alomainy, “Nano-communication for biomedicalapplications: A review on the state-of-the-art from physical layers to novelnetworking concepts,” IEEE Access, vol. 4, pp. 3920–3935, 2016.

[19] F. Dressler and S. Fischer, “Connecting in-body nano communicationwith body area networks: Challenges and opportunities of the internet ofnano things,” Nano Communication Networks, vol. 6, no. 2, pp. 29 – 38,2015.

[20] S. A. Busari, S. Mumtaz, S. Al-Rubaye, and J. Rodriguez, “5Gmillimeter-wave mobile broadband: performance and challenges,” IEEECommunications Magazine, vol. 56, no. 6, pp. 137–143, June 2018.

[21] S. Mumtaz, J. M. Jornet, J. Aulin, W. H. Gerstacker, X. Dong, and B. Ai,“Terahertz communication for vehicular networks,” IEEE Transactions onVehicular Technology, vol. 66, no. 7, pp. 5617–5625, July 2017.

[22] V. Petrov, J. Kokkoniemi, D. Moltchanov, J. Lehtomaki, Y. Koucheryavy,and M. Juntti, “Last meter indoor Terahertz wireless access: Performanceinsights and implementation roadmap,” IEEE Communications Magazine,vol. 56, no. 6, pp. 158–165, June 2018.

[23] K. c. Huang and Z. Wang, “Terahertz terabit wireless communication,”IEEE Microwave Magazine, vol. 12, no. 4, pp. 108–116, June 2011.

[24] C. Han and Y. Chen, “Propagation modeling for wireless communica-tions in the Terahertz band,” IEEE Communications Magazine, vol. 56,no. 6, pp. 96–101, June 2018.

[25] A. A. A. Boulogeorgos, A. Alexiou, T. Merkle, C. Schubert, R. Elschner,A. Katsiotis, P. Stavrianos, D. Kritharidis, P. K. Chartsias, J. Kokkoniemi,M. Juntti, J. Lehtomaki, A. Teixeira, and F. Rodrigues, “Terahertztechnologies to deliver optical network quality of experience in wirelesssystems beyond 5G,” IEEE Communications Magazine, vol. 56, no. 6,pp. 144–151, June 2018.

[26] P. T. Dat, A. Kanno, T. Umezawa, N. Yamamoto, and T. Kawanishi,“Millimeter and Terahertzwave radio-over-fiber for 5G and beyond,” inIEEE Photonics Society Summer Topical Meeting Series (SUM), July2017, pp. 165–166.

[27] V. Petrov, A. Pyattaev, D. Moltchanov, and Y. Koucheryavy, “Terahertzband communications: Applications, research challenges, and standardiza-tion activities,” in 8th International Congress on Ultra Modern Telecom-munications and Control Systems and Workshops (ICUMT), October2016, pp. 183–190.

[28] I. F. Akyildiz and J. M. Jornet, “Realizing ultra-massive MIMO(1024ÃU1024) communication in the (0.06âAS10) Terahertz band,” NanoCommunication Networks, vol. 8, pp. 46 – 54, 2016.

[29] H. J. Song and T. Nagatsuma, “Present and future of Terahertz com-munications,” IEEE Transactions on Terahertz Science and Technology,vol. 1, no. 1, pp. 256–263, September 2011.

[30] I. Hosako, N. Sekine, M. Patrashin, S. Saito, K. Fukunaga, Y. Kasai,P. Baron, T. Seta, J. Mendrok, S. Ochiai, and H. Yasuda, “At the dawnof a new era in Terahertz technology,” Proceedings of the IEEE, vol. 95,no. 8, pp. 1611–1623, August 2007.

[31] M. Hasan, S. Arezoomandan, H. Condori, and B. Sensale-Rodriguez,“Graphene Terahertz devices for communications applications,” NanoCommunication Networks, vol. 10, pp. 68 – 78, 2016.

[32] J. F. Federici, J. Ma, and L. Moeller, “Review of weather impact onoutdoor Terahertz wireless communication links,” Nano CommunicationNetworks, vol. 10, pp. 13 – 26, 2016.

[33] K. Wu, Y. J. Cheng, T. Djerafi, and W. Hong, “Substrate-integratedmillimeter-wave and Terahertz antenna technology,” Proceedings of theIEEE, vol. 100, no. 7, pp. 2219–2232, July 2012.

[34] K. M. S. Huq, J. M. Jornet, W. H. Gerstacker, A. Al-Dulaimi, Z. Zhou,and J. Aulin, “Thz communications for mobile heterogeneous networks,”IEEE Communications Magazine, vol. 56, no. 6, pp. 94–95, June 2018.

[35] C. Han, X. Zhang, and X. Wang, “On medium access control schemesfor wireless networks in the millimeter-wave and terahertz bands,” NanoCommunication Networks, vol. 19, pp. 67 – 80, 2019.

[36] M. Fitch and R. Osiander, “Terahertz waves for communications andsensing,” Johns Hopkins APL Technical Digest, vol. 25, no. 4, pp. 348–355, October 2004.

[37] T. Kurner, “Towards future Terahertz communications systems,” Tera-hertz Science and Technology, vol. 5, pp. 11–17, 01 2012.

[38] C. M. Armstrong, “The truth about Terahertz,” IEEE Spectrum, vol. 49,no. 9, pp. 36–41, September 2012.

[39] S. Balasubramaniam, S. Ben-Yehuda, S. Pautot, A. Jesorka, P. LioâAZ,and Y. Koucheryavy, “A review of experimental opportunities for molec-ular communication,” Nano Communication Networks, vol. 4, no. 2, pp.43 – 52, 2013.

[40] A. Hirata and M. Yaita, “Ultrafast Terahertz wireless communicationstechnologies,” IEEE Transactions on Terahertz Science and Technology,vol. 5, no. 6, pp. 1128–1132, November 2015.

[41] N. Khalid, T. Yilmaz, and O. B. Akan, “Energy-efficient modulation andphysical layer design for low Terahertz band communication channel in5G femtocell internet of things,” Ad Hoc Networks, vol. 79, pp. 63 – 71,2018.

[42] F. Lemic, S. Abadal, W. Tavernier, P. Stroobant, D. Colle,E. AlarcÃsn, J. Marquez-Barja, and J. Famaey, “Survey on terahertznanocommunication and networking: A top-down perspective,” 2019.[Online]. Available: https://arxiv.org/abs/1909.05703

[43] T. S. Rappaport, Y. Xing, O. Kanhere, S. Ju, A. Madanayake, S. Mandal,A. Alkhateeb, and G. C. Trichopoulos, “Wireless communications andapplications above 100 ghz: Opportunities and challenges for 6g andbeyond,” IEEE Access, vol. 7, pp. 78 729–78 757, 2019.

[44] H. Elayan, O. Amin, B. Shihada, R. M. Shubair, and M.-S. Alouini,“Terahertz band: The last piece of rf spectrum puzzle for communicationsystems,” arXiv e-prints, p. arXiv:1907.05043, Jul 2019.

[45] K. Tekbiyik, A. R. Ekti, G. K. Kurt, and A. Gorcin,“Terahertz band communication systems: Challenges, novelties andstandardization efforts,” Physical Communication, vol. 35, p. 100700,

42

2019. [Online]. Available: http://www.sciencedirect.com/science/article/pii/S1874490718307766

[46] K. K. O, W. Choi, Q. Zhong, N. Sharma, Y. Zhang, R. Han, Z. Ahmad,D. Kim, S. Kshattry, I. R. Medvedev, D. J. Lary, H. Nam, P. Raskin, andI. Kim, “Opening terahertz for everyday applications,” IEEE Communi-cations Magazine, vol. 57, no. 8, pp. 70–76, August 2019.

[47] K. M. S. Huq, S. A. Busari, J. Rodriguez, V. Frascolla, W. Bazzi, andD. C. Sicker, “Terahertz-enabled wireless system for beyond-5g ultra-fastnetworks: A brief survey,” IEEE Network, vol. 33, no. 4, pp. 89–95, July2019.

[48] Z. Chen, X. Ma, B. Zhang, Y. Zhang, Z. Niu, N. Kuang, W. Chen, L. Li,and S. Li, “A survey on terahertz communications,” China Communica-tions, vol. 16, no. 2, pp. 1–35, February 2019.

[49] I. F. Akyildiz, F. Brunetti, and C. BlÃazquez, “Nanonetworks: A newcommunication paradigm,” Computer Networks, vol. 52, no. 12, pp. 2260– 2279, 2008.

[50] I. F. Akyildiz and J. M. Jornet, “Electromagnetic wireless nanosensornetworks,” Nano Communication Networks, vol. 1, no. 1, pp. 3 – 19,2010.

[51] J. M. Jornet and I. F. Akyildiz, “The internet of multimedia nano-thingsin the terahertz band,” in 18th European Wireless Conference, April 2012,pp. 1–8.

[52] S. V. Sivapriya and D. Sridharan, “A comprehensive review on MACprotocol of wireless nanosensor networks for intrabody application,”International Journal for Innovative Research in Science and Technology,vol. 4, no. 2, pp. 83–89, July 2017.

[53] X. Wang, L. Kong, F. Kong, F. Qiu, M. Xia, S. Arnon, and G. Chen,“Millimeter wave communication: A comprehensive survey,” IEEE Com-munications Surveys & Tutorials, vol. 20, no. 3, pp. 1616–1653, thirdquar-ter 2018.

[54] J. Ma, “Terahertz wireless communication through atmospheric atmo-spheric turbulence and rain,” 2015.

[55] A. S. Cacciapuoti, K. Sankhe, M. Caleffi, and K. R. Chowdhury,“Beyond 5G: Terahertz-based Medium Access Protocol for mobile het-erogeneous networks,” IEEE Communications Magazine, vol. 56, no. 6,pp. 110–115, June 2018.

[56] I. F. Akyildiz, D. M. Gutierrez-Estevez, R. Balakrishnan, andE. Chavarria-Reyes, “LTE-Advanced and the evolution to Beyond 4G(B4G) systems,” Physical Communication, vol. 10, pp. 31 – 60, 2014.

[57] B. Glushko, D. Kin, and A. Shar, “Gigabit optical wireless communica-tion system for personal area networking,” Optical Memory and NeuralNetworks, vol. 22, no. 2, pp. 73–80, April 2013.

[58] X. Li, J. Vucic, V. Jungnickel, and J. Armstrong, “On the capacity ofintensity-modulated direct-detection systems and the information rate ofACO-OFDM for indoor optical wireless applications,” IEEE Transactionson Communications, vol. 60, no. 3, pp. 799–809, March 2012.

[59] E. Ciaramella, Y. Arimoto, G. Contestabile, M. Presi, A. D’Errico,V. Guarino, and M. Matsumoto, “1.28 terabit/s (32x40 gbit/s) WDMtransmission system for free space optical communications,” IEEE Jour-nal on Selected Areas in Communications, vol. 27, no. 9, pp. 1639–1645,December 2009.

[60] E. Calvanese Strinati, S. Barbarossa, J. L. Gonzalez-Jimenez, D. Ktenas,N. Cassiau, L. Maret, and C. Dehos, “6g: The next frontier: Fromholographic messaging to artificial intelligence using subterahertz andvisible light communication,” IEEE Vehicular Technology Magazine,vol. 14, no. 3, pp. 42–50, Sep. 2019.

[61] M. Fujishima, S. Amakawa, K. Takano, K. Katayama, and T. Yoshida,“Tehrahertz CMOS design for low-power and high-speed wireless com-munication,” IEICE Transactions, vol. 98-C, pp. 1091–1104, 2015.

[62] I. Kallfass, I. Dan, S. Rey, P. Harati, J. Antes, A. Tessmann, S. Wagner,M. Kuri, R. Weber, H. Massler, A. Leuther, T. Merkle, and T. Kürner,“Towards MMIC-based 300ghz indoor wireless communication systems,”IEICE Transactions, vol. 98-C, pp. 1081–1090, 2015.

[63] X. Yu, S. Jia, H. Hu, P. Guan, M. Galili, T. Morioka, P. U. Jepsen,and L. K. OxenlÃÿwe, “Terahertz photonics-wireless transmission of 160Gbit/s bitrate,” in 21st OptoElectronics and Communications Conference(OECC) held jointly with International Conference on Photonics inSwitching (PS), July 2016, pp. 1–3.

[64] X. Pang, S. Jia, O. Ozolins, X. Yu, H. Hu, L. Marcon, P. Guan, F. D.Ros, S. Popov, G. Jacobsen, M. Galili, T. Morioka, D. Zibar, and L. K.Oxenkwe, “260 gbits photonic-wireless link in the Terahertz band,” inIEEE Photonics Conference (IPC), Oct 2016, pp. 1–2.

[65] P. H. Pathak, X. Feng, P. Hu, and P. Mohapatra, “Visible light commu-nication, networking, and sensing: A survey, potential and challenges,”IEEE Communications Surveys & Tutorials, vol. 17, no. 4, pp. 2047–2077, Fourthquarter 2015.

[66] S. Rajagopal, R. D. Roberts, and S. Lim, “IEEE 802.15.7 visiblelight communication: modulation schemes and dimming support,” IEEECommunications Magazine, vol. 50, no. 3, pp. 72–82, March 2012.

[67] H. S. Ghadikolaei and C. Fischione, “The transitional behavior ofinterference in millimeter wave networks and its impact on medium accesscontrol,” IEEE Transactions on Communications, vol. 64, pp. 723–740,2015.

[68] L. Marnat, L. Dussopt, V. Puyal, A. Siligaris, F. Hameau, A. Larie,and C. Dehos, “V-band transceiver modules with integrated antennas andphased arrays for mmwave access in 5g mobile networks,” in 2017 11thEuropean Conference on Antennas and Propagation (EUCAP), March2017, pp. 2786–2790.

[69] W. Roh, J. Seol, J. Park, B. Lee, J. Lee, Y. Kim, J. Cho, K. Cheun, andF. Aryanfar, “Millimeter-wave beamforming as an enabling technology for5g cellular communications: theoretical feasibility and prototype results,”IEEE Communications Magazine, vol. 52, no. 2, pp. 106–113, February2014.

[70] H. Jalili and O. Momeni, “Scalable wideband and wide-angle beamsteering mm-wave/thz radiator and phased arrays in silicon,” in 2018Asia-Pacific Microwave Conference (APMC), Nov 2018, pp. 1157–1159.

[71] I. A. Hemadeh, K. Satyanarayana, M. El-Hajjar, and L. Hanzo,“Millimeter-wave communications: Physical channel models, design con-siderations, antenna constructions, and link-budget,” IEEE Communica-tions Surveys Tutorials, vol. 20, no. 2, pp. 870–913, Secondquarter 2018.

[72] V. K. Chinni, M. Zezaoui, C. Coinon, X. Wallart, E. Pevtavit, J. F.Larnpin, P. Szriftgiser, M. Zaknoune, and G. Ducournau, “Indoor 100gbit/s thz data link in the 300 ghz band using fast photodiodes,” in 201825th International Conference on Telecommunications (ICT), June 2018,pp. 288–290.

[73] N. Rupasinghe, Y. Kakishima, I. Guvenc, K. Kitao, and T. Imai,“Geometry performance for 5g mmwave cellular networks,” in 2016International Symposium on Antennas and Propagation (ISAP), Oct 2016,pp. 874–875.

[74] M. Giordani, M. Polese, A. Roy, D. Castor, and M. Zorzi, “A tutorialon beam management for 3gpp nr at mmwave frequencies,” IEEE Com-munications Surveys Tutorials, vol. 21, no. 1, pp. 173–196, Firstquarter2019.

[75] H. Wang, P. Zhang, J. Li, and X. You, “Radio propagation and wire-less coverage of lsaa-based 5g millimeter-wave mobile communicationsystems,” China Communications, vol. 16, no. 5, pp. 1–18, May 2019.

[76] M. Mezzavilla, M. Zhang, M. Polese, R. Ford, S. Dutta, S. Rangan, andM. Zorzi, “End-to-end simulation of 5g mmwave networks,” IEEE Com-munications Surveys Tutorials, vol. 20, no. 3, pp. 2237–2263, thirdquarter2018.

[77] OPENAIRINTERFACE, https://www.openairinterface.org/?page_id=1831, accessed: 26-10-2019.

[78] Z. Hossain, Q. Xia, and J. M. Jornet, “Terasim: An ns-3 extension tosimulate Terahertz band communication networks,” Nano CommunicationNetworks, vol. 17, pp. 36 – 44, 2018.

[79] IEEE, “IEEE 802.15.3d: Technical requirement document,”techreport, 2015. [Online]. Available: https://mentor.ieee.org/802.15/dcn/14/15-14-0309-20-003d-technical-requirements-document.docx

[80] T. Yilmaz and O. B. Akan, “Utilizing terahertz band for local and per-sonal area wireless communication systems,” in IEEE 19th InternationalWorkshop on Computer Aided Modeling and Design of CommunicationLinks and Networks (CAMAD), December 2014, pp. 330–334.

[81] C. Wang, B. Lu, C. Lin, Q. Chen, L. Miao, X. Deng, and J. Zhang, “0.34-thz wireless link based on high-order modulation for future wireless localarea network applications,” IEEE Transactions on Terahertz Science andTechnology, vol. 4, no. 1, pp. 75–85, January 2014.

[82] “IEEE standard for high data rate wireless multi-media networks–amendment 2: 100 gb/s wireless switched point-to-point physical layer,”IEEE Std 802.15.3d-2017 (Amendment to IEEE Std 802.15.3-2016 asamended by IEEE Std 802.15.3e-2017), pp. 1–55, Oct 2017.

[83] H. Song, H. Hamada, and M. Yaita, “Prototype of KIOSK data down-loading system at 300 ghz: Design, technical feasibility, and results,”IEEE Communications Magazine, vol. 56, no. 6, pp. 130–136, June 2018.

[84] L. You, Z. Ren, C. Chen, and Y.-H. Lv, “An improved high throughputand low delay access protocol for Terahertz wireless personal areanetworks,” Journal of Computers, vol. 28, no. 3, pp. 147–158, June 2017.

[85] V. Petrov, D. Moltchanov, and Y. Koucheryavy, “Applicability as-sessment of terahertz information showers for next-generation wirelessnetworks,” in IEEE International Conference on Communications (ICC),May 2016, pp. 1–7.

[86] IEEE, “IEEE 802.15.3d: Application requirement document,” Tech.Rep., 2015.

43

[87] P. T. Dat, A. Kanno, N. Yamamoto, and T. Kawanishi, “Radio-over-fiber-based seamless fiber-wireless convergence for small cell and linearcell networks,” in 2018 Optical Fiber Communications Conference andExposition (OFC), March 2018, pp. 1–3.

[88] M. T. Barros, R. Mullins, and S. Balasubramaniam, “Integrated tera-hertz communication with reflectors for 5g small-cell networks,” IEEETransactions on Vehicular Technology, vol. 66, no. 7, pp. 5647–5657,July 2017.

[89] V. Petrov, D. Moltchanov, and Y. Koucheryavy, “Interference andsinr in dense Terahertz networks,” in IEEE 82nd Vehicular TechnologyConference (VTC2015-Fall), September 2015, pp. 1–5.

[90] A. S. Cacciapuoti, S. Ramanathan, K. R. Chowdhury, and M. Caleffi,“Software-defined network controlled switching between millimeter waveand Terahertz small cells,” CoRR, vol. abs/1702.02775, 2017.

[91] K. Guan, G. Li, T. KÃijrner, A. F. Molisch, B. Peng, R. He, B. Hui,J. Kim, and Z. Zhong, “On millimeter wave and THz mobile radio channelfor smart rail mobility,” IEEE Transactions on Vehicular Technology,vol. 66, no. 7, pp. 5658–5674, July 2017.

[92] C. Zhang, K. Ota, J. Jia, and M. Dong, “Breaking the blockage for bigdata transmission: Gigabit road communication in autonomous vehicles,”IEEE Communications Magazine, vol. 56, no. 6, pp. 152–157, June 2018.

[93] P. Kumari, J. Choi, N. GonzÃalez-Prelcic, and R. W. Heath, “Ieee802.11ad-based radar: An approach to joint vehicular communication-radar system,” IEEE Transactions on Vehicular Technology, vol. 67, no. 4,pp. 3012–3027, April 2018.

[94] V. Petrov, G. Fodor, J. Kokkoniemi, D. Moltchanov, J. Lehtomäki,S. Andreev, Y. Koucheryavy, M. J. Juntti, and M. Valkama, “On unifiedvehicular communications and radar sensing in millimeter-wave and lowterahertz bands,” CoRR, vol. abs/1901.06980, 2019. [Online]. Available:http://arxiv.org/abs/1901.06980

[95] D. L. Woolard, J. O. Jensen, R. J. Hwu, and M. S. Shur, TerahertzScience and Technology for Military and Security Applications. WORLDSCIENTIFIC, 2007.

[96] S. Sonmez and S. Ergun, “Terahertz technology for military applica-tions,” Journal of Management and Information Science, vol. 3, no. 1,pp. 13 – 16, 2015.

[97] K. Iwaszczuk, P. Jepsen, and H. Heiselberg, “Terahertz technology fordefense and security-related applications,” Ph.D. dissertation, 2012.

[98] S. U. Hwu, K. B. deSilva, and C. T. Jih, “Terahertz wireless systems forspace applications,” in IEEE Sensors Applications Symposium Proceed-ings, February 2013, pp. 171–175.

[99] S. Dong, Z. Zhu, and Y. Wang, “Advances of terahertz researchand Terahertz satellite communications,” in International Conference onElectronics, Communications and Control (ICECC), September 2011, pp.4122–4125.

[100] T. M. Narytnyk, “Possibilities of using Terahertz band radio commu-nication channels for super high rate backhaul,” Telecommunications andRadio Engineering, vol. 73, no. 15, pp. 1361–1371, 2014.

[101] L. Felicetti, M. Femminella, G. Reali, and P. LiÚ, “Applications ofmolecular communications to medicine: A survey,” Nano CommunicationNetworks, vol. 7, pp. 27 – 45, 2016.

[102] E. Zarepour, M. Hassan, C. T. Chou, A. A. Adesina, and M. E.Warkiani, “Reliability analysis of time-varying wireless nanoscale sensornetworks,” in 2015 IEEE 15th International Conference on Nanotechnol-ogy (IEEE-NANO), July 2015, pp. 63–68.

[103] Z. L. Wang, “Towards self-powered nanosystems: From nanogeneratorsto nanopiezotronics,” Advanced Functional Materials, vol. 18, no. 22, pp.3553–3567.

[104] S. Mathanker, “Terahertz applications in food and agriculture: Areview,” Transactions of the ASABE (American Society of Agriculturaland Biological Engineers), vol. 56, pp. 1213–1226, April 2013.

[105] A. Zahid, K. Yang, H. Heidari, C. Li, M. A. Imran, A. Alomainy, andQ. H. Abbasi, “Terahertz characterisation of living plant leaves for qualityof life assessment applications,” in 2018 Baltic URSI Symposium (URSI),May 2018, pp. 117–120.

[106] A. Afsharinejad, A. Davy, and B. Jennings, “Dynamic channel alloca-tion in electromagnetic nanonetworks for high resolution monitoring ofplants,” Nano Communication Networks, vol. 7, pp. 2 – 16, 2016.

[107] N. Khalid, N. A. Abbasi, and O. B. Akan, “300 ghz broadbandtransceiver design for low-thz band wireless communications in indoorinternet of things,” in 2017 IEEE International Conference on Internetof Things (iThings) and IEEE Green Computing and Communications(GreenCom) and IEEE Cyber, Physical and Social Computing (CPSCom)and IEEE Smart Data (SmartData), June 2017, pp. 770–775.

[108] V. Petrov, D. Moltchanov, M. Komar, A. Antonov, P. Kustarev,S. Rakheja, and Y. Koucheryavy, “Terahertz band intra-chip communica-

tions: Can wireless links scale modern x86 cpus?” IEEE Access, vol. 5,pp. 6095–6109, 2017.

[109] O. Yalgashev, “Towards nanoscale interconnect for system-on-chip,”Ph.D. dissertation, UniversitÃl’ de Technologie de Belfort-Montbeliard,2018.

[110] S. Kim and A. ZajiÄG, “Characterization of 300-ghz wireless chan-nel on a computer motherboard,” IEEE Transactions on Antennas andPropagation, vol. 64, no. 12, pp. 5411–5423, Dec 2016.

[111] “Requirements on wireless backhauling andfronthauling, ieee 802.15-08-0336-00-0thz,” November,2013. [Online]. Available: https://mentor.ieee.org/802.15/dcn/13/15-13-0636-01-0thz-requirements-for-wireless-backhauling-fronthauling.pdf

[112] S. Ahearne, N. O. Mahony, N. Boujnah, S. Ghafoor, L. G. Guerrero,and C. Renaud, “Integrating thz wireless communication links in a datacentre network,” in IEEE 5G Worlf Forum, October 2019.

[113] S. Koenig, D. Lopez-Diaz, J. Antes, F. Boes, R. Henneberger,A. Leuther, A. Tessmann, R. Schmogrow, D. Hillerkuss, R. Palmer,T. Zwick, C. Koos, W. Freude, O. Ambacher, J. Leuthold, and I. Kallfass,“Wireless sub-thz communication system with high data rate,” NaturePhotonics, vol. 7, pp. 977–981, October 2013.

[114] Y. Katayama, K. Takano, Y. Kohda, N. Ohba, and D. Nakano, “Wirelessdata center networking with steered-beam mmwave links,” in IEEEWireless Communications and Networking Conference, March 2011, pp.2179–2184.

[115] S. Mollahasani and E. Onur, “Evaluation of terahertz channel in datacenters,” in NOMS 2016 - 2016 IEEE/IFIP Network Operations andManagement Symposium, April 2016, pp. 727–730.

[116] K. Wu, J. Xiao, and L. M. Ni, “Rethinking the architecture design ofdata center networks,” Frontiers of Computer Science, vol. 6, no. 5, pp.596–603, October 2012.

[117] NTT, http://www.ntt.co.jp/news2016/1605e/160526a.html, accessed:09-01-2019.

[118] A. D. Angelica, GoogleâAZs self-driving car gathers nearly1 GBps, 2013. [Online]. Available: http://www.kurzweilai.net/googles-self-driving-car-gathers-nearly-1-gbsec.

[119] SAS, Are you ready for your smart car? [Online].Available: http://www.sas.com/enus/insights/articles/big-data/the-internet-of-things-and-connected-cars.html

[120] A. Kanno, “Fiber-wireless signal transport by Terahertz waves,” inInternational Conference on Advanced Technologies for Communications(ATC 2014), Oct 2014, pp. 766–769.

[121] A. Greenberg, J. Hamilton, D. A. Maltz, and P. Patel, “The cost of acloud: Research problems in data center networks,” SIGCOMM Comput.Commun. Rev., vol. 39, no. 1, pp. 68–73, December 2008.

[122] Y.-M. Ding, S. Gao, X. Shi, and H. Wu, “Analysis of inter-satellite ter-ahertz communication link,” in 3rd International Conference on WirelessCommunication and Sensor Networks (WCSN). Atlantis Press, 2016.

[123] W. Dou, L. Zhang, H. Meng, and Z. Wang, “Tracking antennas forinter-satellite communications at sub-millimeter wavelengths,” in Pro-ceedings of 3rd Asia-Pacific Conference on Antennas and Propagation,July 2014, pp. 1149–1152.

[124] T. Yilmaz and O. B. Akan, “On the use of low Terahertz band for 5Gindoor mobile networks,” Computers and Electrical Engineering, vol. 48,pp. 164 – 173, 2015.

[125] THOR, https://thorproject.eu/, accessed: 01-04-2019.[126] A. A. Boulogeorgos, A. Alexiou, D. Kritharidis, A. Katsiotis, G. D.

Ntouni, J. Kokkoniemi, J. Lethtomaki, M. J. Juntti, D. Yankova,A. Mokhtar, J. Point, J. Machado, R. Elschner, C. Schubert, T. Merkle,R. Ferreira, F. Rodrigues, and J. Lima, “Wireless terahertz systemarchitectures for networks beyond 5g,” CoRR, vol. abs/1810.12260,2018. [Online]. Available: http://arxiv.org/abs/1810.12260

[127] S. Mohrehkesh, M. C. Weigle, and S. K. Das, “DRIH-MAC: Adistributed receiver-initiated harvesting-aware MAC for nanonetworks,”IEEE Transactions on Molecular, Biological and Multi-Scale Communi-cations, vol. 1, no. 1, pp. 97–110, March 2015.

[128] G. K. Walia, D. K. K. Randhawa, and K. S. Malhi, “A brief survey onmolecular communications in nanonetworks,” in International Confer-ence on Computational Techniques in Information and CommunicationTechnologies (ICCTICT), March 2016, pp. 343–348.

[129] T. Qiu, T.-C. Lee, A. G. Mark, K. I. Morozov, R. Münster, O. Mierka,S. Turek, A. M. Leshansky, and P. Fischer, “Swimming by reciprocalmotion at low reynolds number,” in Nature communications, 2014.

[130] C. Cvetkovic, R. Raman, V. Chan, B. J. Williams, M. Tolish, P. Bajaj,M. S. Sakar, H. H. Asada, M. T. A. Saif, and R. Bashir, “Three-dimensionally printed biological machines powered by skeletal muscle,”

44

Proceedings of the National Academy of Sciences, vol. 111, no. 28, pp.10 125–10 130, 2014.

[131] S. Yim and M. Sitti, “Design and rolling locomotion of a magneticallyactuated soft capsule endoscope,” IEEE Transactions on Robotics, vol. 28,no. 1, pp. 183–194, February 2012.

[132] F. Qiu, S. Fujita, R. Mhanna, L. Zhang, B. R. Simona, and B. J.Nelson, “Magnetic helical microswimmers functionalized with lipoplexesfor targeted gene delivery,” Advanced Functional Materials, vol. 25,no. 11, pp. 1666–1671, 2015.

[133] S. Abadal, E. AlarcÃsn, A. Cabellos-Aparicio, M. C. Lemme, andM. Nemirovsky, “Graphene-enabled wireless communication for mas-sive multicore architectures,” IEEE Communications Magazine, vol. 51,no. 11, pp. 137–143, November 2013.

[134] A. Amro, M. Ilchenko, V. Kalinin, and O. Turabi, “Sub-terahertzlow power uwb communication link for wpan,” in IISTE Network andComplex Systems, 2012.

[135] C. Chaccour, R. Amer, B. Zhou, and W. Saad, “On the reliabilityof wireless virtual reality at terahertz (thz) frequencies,” CoRR, vol.abs/1905.07656, 2019. [Online]. Available: http://arxiv.org/abs/1905.07656

[136] D. Arifler, “Link layer modeling of bio-inspired communication innanonetworks,” Nano Communication Networks, vol. 2, no. 4, pp. 223– 229, 2011.

[137] J. M. Jornet, J. C. Pujol, and J. S. Pareta, “PHLAME: A physical layeraware MAC protocol for electromagnetic nanonetworks in the Terahertzband,” Nano Communication Networks, vol. 3, no. 1, pp. 74 – 81, 2012.

[138] V. Srikanth, S. Chaluvadi, Sandeep, Vani, and Venkatesh, “Energyefficient, scalable and reliable MAC protocol for electromagnetic commu-nication among nano devices,” International Journal of Distributed andParallel Systems (IJDPS), vol. 03, no. 1, pp. 249–256, January 2012.

[139] P. Wang, J. M. Jornet, M. A. Malik, N. Akkari, and I. F. Akyildiz, “En-ergy and spectrum-aware MAC protocol for perpetual wireless nanosensornetworks in the Terahertz band,” Ad Hoc Networks, vol. 11, no. 8, pp.2541 – 2555, 2013.

[140] Z. Ren, Y. N. Cao, S. Peng, and H. J. Lei, “A MAC protocol forTerahertz ultra-high data-rate wireless networks,” in Mechanical Engi-neering, Industrial Electronics and Information Technology Applicationsin Industry, ser. Applied Mechanics and Materials, vol. 427. Trans TechPublications, 12 2013, pp. 2864–2869.

[141] G. Piro, L. A. Grieco, G. Boggia, and P. Camarda, “Nano-sim:Simulating electromagnetic-based nanonetworks in the network simulator3,” in Proceedings of the 6th International ICST Conference on SimulationTools and Techniques, ser. SimuTools ’13. ICST, Brussels, Belgium,Belgium: ICST (Institute for Computer Sciences, Social-Informatics andTelecommunications Engineering), 2013, pp. 203–210.

[142] Z. Ren, Y.-N. Cao, X. Zhou, Y. Zheng, and Q. bin Chen, “Novel MACprotocol for Terahertz ultra-high data-rate wireless networks,” The Journalof China Universities of Posts and Telecommunications, vol. 20, no. 6,pp. 69 – 76, 2013.

[143] J. Lin and M. A. Weitnauer, “Pulse-level beam-switching MAC withenergy control in picocell Terahertz networks,” in IEEE Global Commu-nications Conference (GLOBECOM), December 2014, pp. 4460–4465.

[144] S. Mohrehkesh and M. C. Weigle, “Optimizing energy consumptionin Terahertz band nanonetworks,” IEEE Journal on Selected Areas inCommunications, vol. 32, no. 12, pp. 2432–2441, December 2014.

[145] E. Zarepour, M. Hassan, C. T. Chou, and A. A. Adesina, “Frequencyhopping strategies for improving Terahertz sensor network performanceover composition varying channels,” in Proceeding of IEEE InternationalSymposium on a World of Wireless, Mobile and Multimedia Networks2014, June 2014, pp. 1–9.

[146] S. Mohrehkesh and M. C. Weigle, “RIH-MAC: Receiver-initiatedharvesting-aware MAC for nanonetworks,” in Proceedings of ACM TheFirst Annual International Conference on Nanoscale Computing andCommunication, ser. NANOCOM’ 14. New York, NY, USA: ACM,2014, pp. 1–6.

[147] Q. Xia, Z. Hossain, M. Medley, and J. M. Jornet, “A link-layersynchronization and Medium Access Control protocol for Terahertz bandcommunication networks,” in IEEE Global Communications Conference(GLOBECOM), December 2015, pp. 1–7.

[148] S. DâAZOro, L. Galluccio, G. Morabito, and S. Palazzo, “A timingchannel-based MAC protocol for energy-efficient nanonetworks,” NanoCommunication Networks, vol. 6, no. 2, pp. 39 – 50, 2015.

[149] N. Akkari, J. M. Jornet, P. Wang, E. Fadel, L. Elrefaei, M. G. Malik,S. Almasri, and I. F. Akyildiz, “Joint physical and link layer error controlanalysis for nanonetworks in the Terahertz band,” Wirel. Netw., vol. 22,no. 4, pp. 1221–1233, May 2016.

[150] S. J. Lee, C. A. Jung, K. Choi, and S. Kim, “Design of wirelessnanosensor networks for intrabody application,” International Journal ofDistributed Sensor Networks, vol. 11, no. 7, pp. 1–12, 2015.

[151] N. Akkari, P. Wang, J. M. Jornet, E. Fadel, L. Elrefaei, M. G. A.Malik, S. Almasri, and I. F. Akyildiz, “Distributed timely throughputoptimal scheduling for the internet of nano-things,” IEEE Internet ofThings Journal, vol. 3, no. 6, pp. 1202–1212, December 2016.

[152] X.-W. Yao and J. M. Jornet, “TAB-MAC: Assisted beamforming MACprotocol for Terahertz communication networks,” Nano CommunicationNetworks, vol. 9, pp. 36 – 42, 2016.

[153] R. Alsheikh, N. Akkari, and E. Fadel, “Grid based energy-aware MACprotocol for wireless nanosensor network,” in 8th IFIP International Con-ference on New Technologies, Mobility and Security (NTMS), November2016, pp. 1–5.

[154] A. Tsioliaridou, C. Liaskos, S. Ioannidis, and A. Pitsillides,“Lightweight, self-tuning data dissemination for dense nanonetworks,”Nano Communication Networks, vol. 8, pp. 2 – 15, 2016.

[155] P. Johari and J. M. Jornet, “Packet size optimization for wirelessnanosensor networks in the Terahertz band,” in IEEE InternationalConference on Communications (ICC), May 2016, pp. 1–6.

[156] R. Alsheikh, N. Akkari, and E. Fadel, “MAC protocols for wirelessnano-sensor networks: Performance analysis and design guidelines,” inSixth International Conference on Digital Information Processing andCommunications (ICDIPC), April 2016, pp. 129–134.

[157] C. Jianling, W. Min, C. Cong, and R. Zhi, “High-throughput low-delayMAC protocol for Terahertz ultra-high data-rate wireless networks,” TheJournal of China Universities of Posts and Telecommunications, vol. 23,no. 4, pp. 17 – 24, 2016.

[158] C. Han, W. Tong, and X.-W. Yao, “MA-ADM: A memory-assistedangular-division multiplexing MAC protocol in Terahertz communicationnetworks,” Nano Communication Networks, vol. 13, pp. 51 – 59, 2017.

[159] L. Lopacinski, M. Brzozowski, and R. Kraemer, “Data link layerconsiderations for future 100 Gbps Terahertz band transceivers,” Wirelesscommunications and mobile computing, 2017.

[160] W. Tong and C. Han, “MRA-MAC: A multi-radio assisted mediumaccess control in terahertz communication networks,” in IEEE GlobalCommunications Conference (GLOBECOM), December 2017, pp. 1–6.

[161] Q. Li, X.-W. Yao, and C.-C. Wang, “RBMP: A relay-based MACprotocol for nanonetworks in the Terahertz band,” in Proceedings ofthe 4th ACM International Conference on Nanoscale Computing andCommunication, ser. NanoCom. New York, NY, USA: ACM, 2017,pp. 1–2.

[162] H. Yu, B. Ng, and W. K. G. Seah, “Pulse arrival scheduling for nanonet-works under limited IoT access bandwidth,” in IEEE 42nd Conferenceon Local Computer Networks (LCN), October 2017, pp. 18–26.

[163] Q. Xia and J. M. Jornet, “Cross-layer analysis of optimal relayingstrategies for Terahertz band communication networks,” in IEEE 13thInternational Conference on Wireless and Mobile Computing, Networkingand Communications (WiMob), October 2017, pp. 1–8.

[164] N. Rikhtegar, M. Keshtgari, and Z. Ronaghi, “EEWNSN: Energyefficient wireless nano sensor network MAC protocol for communicationsin the Terahertz band,” Wireless Personal Communications, vol. 97, no. 1,pp. 521–537, November 2017.

[165] Z. Li, L. Guan, C. Li, and A. Radwan, “A secure intelligent spectrumcontrol strategy for future Terahertz mobile heterogeneous networks,”IEEE Communications Magazine, vol. 56, no. 6, pp. 116–123, June 2018.

[166] X. Yao, C. Wang, W. Wang, and J. M. Jornet, “On the achievablethroughput of energy-harvesting nanonetworks in the Terahertz band,”IEEE Sensors Journal, vol. 18, no. 2, pp. 902–912, January 2018.

[167] S.-J. Lee, H. Choi, and S. U. Kim, “Slotted CSMA/CA basedenergy efficient MAC protocol design in nanonetworks,” CoRR, vol.abs/1803.00900, 2018.

[168] S. E. Hosseininejad, S. Abadal, M. Neshat, R. Faraji-Dana, M. C.Lemme, C. Suessmeier, P. H. Bolà var, E. AlarcÃsn, and A. Cabellos-Aparicio, “MAC-oriented programmable terahertz PHY via graphene-based yagi-uda antennas,” in IEEE Wireless Communications and Net-working Conference (WCNC), April 2018, pp. 1–6.

[169] S. Salous, Radio Propagation Measurement and Channel Modelling,1st ed. Wiley Publishing, 2013.

[170] J. M. Jornet and I. F. Akyildiz, “Channel modeling and capacityanalysis for electromagnetic wireless nanonetworks in the Terahertzband,” IEEE Transactions on Wireless Communications, vol. 10, no. 10,pp. 3211–3221, October 2011.

[171] I. Gordon, L. Rothman, C. Hill, R. Kochanov, Y. Tan, P. Bernath,M. Birk, V. Boudon, A. Campargue, K. Chance, B. Drouin, J.-M.Flaud, R. Gamache, J. Hodges, D. Jacquemart, V. Perevalov, A. Perrin,K. Shine, M.-A. Smith, J. Tennyson, G. Toon, H. Tran, V. Tyuterev,

45

A. Barbe, A. CsÃaszÃar, V. Devi, T. Furtenbacher, J. Harrison, J.-M.Hartmann, A. Jolly, T. Johnson, T. Karman, I. Kleiner, A. Kyuberis,J. Loos, O. Lyulin, S. Massie, S. Mikhailenko, N. Moazzen-Ahmadi,H. MÃijller, O. Naumenko, A. Nikitin, O. Polyansky, M. Rey, M. Rotger,S. Sharpe, K. Sung, E. Starikova, S. Tashkun, J. V. Auwera, G. Wagner,J. Wilzewski, P. WcisÅCo, S. Yu, and E. Zak, “The HITRAN2016molecular spectroscopic database,” Journal of Quantitative Spectroscopyand Radiative Transfer, vol. 203, pp. 3 – 69, 2017, hITRAN2016 SpecialIssue.

[172] H. Cui, J. Yao, and C. Wan, “The study on Terahertz wave propagationfeature in atmosphere,” Journal of Physics: Conference Series, vol. 276,no. 1, pp. 1–7, 2011.

[173] H. Song, K. Ajito, A. Wakatsuki, Y. Muramoto, N. Kukutsu, Y. Kado,and T. Nagatsuma, “Terahertz wireless communication link at 300 ghz,”in IEEE International Topical Meeting on Microwave Photonics, October2010, pp. 42–45.

[174] S. Priebe, C. Jastrow, M. Jacob, T. Kleine-Ostmann, T. Schrader, andT. Kurner, “Channel and propagation measurements at 300 GHz,” IEEETransactions on Antennas and Propagation, vol. 59, no. 5, pp. 1688–1698, May 2011.

[175] S. Priebe and T. Kurner, “Stochastic modeling of Terahertz indoor radiochannels,” IEEE Transactions on Wireless Communications, vol. 12, no. 9,pp. 4445–4455, September 2013.

[176] S. Priebe, M. Kannicht, M. Jacob, and T. KÃijrner, “Ultra broadbandindoor channel measurements and calibrated ray tracing propagationmodeling at terahertz frequencies,” Journal of Communications andNetworks, vol. 15, no. 6, pp. 547–558, December 2013.

[177] K. Tsujimura, K. Umebayashi, J. Kokkoniemi, J. LehtomÃd’ki, andY. Suzuki, “A causal channel model for the Terahertz band,” IEEETransactions on Terahertz Science and Technology, vol. 8, no. 1, pp.52–62, January 2018.

[178] A. Afsharinejad, A. Davy, B. Jennings, S. Rasmann, and C. Bren-nan, “A path-loss model incorporating shadowing for Terahertz bandpropagation in vegetation,” in IEEE Global Communications Conference(GLOBECOM), December 2015, pp. 1–6.

[179] J. Kokkoniemi, J. LehtomÃd’ki, and M. Juntti, “A discussion onmolecular absorption noise in the Terahertz band,” Nano CommunicationNetworks, vol. 8, pp. 35 – 45, 2016.

[180] ITU, “Recommendation ITU-R P.525-3: Calculation of free-space attenuation,” Tech. Rep., 2016, accessed: 01-04-2019.[Online]. Available: https://www.itu.int/dms_pubrec/itu-r/rec/p/R-REC-P.525-3-201611-I!!PDF-E.pdf

[181] ——, “Recommendation ITU-R P.676-10: Attenuation byatmospheric gases,” Tech. Rep., 2013, accessed: 01-04-2019.[Online]. Available: https://www.itu.int/dms_pubrec/itu-r/rec/p/R-REC-P.676-10-201309-S!!PDF-E.pdf

[182] S. Bhardwaj, N. Nahar, and J. L. Volakis, “Link budget analysis for 350GHz communication link,” in US National Committee of URSI NationalRadio Science Meeting (USNC-URSI NRSM), January 2013, pp. 1–1.

[183] D. A. B. Miller, Quantum Mechanics for Scientists and Engineers.Cambridge University Press, 2008.

[184] ITU, “Recommendation ITU-R P.676-11: Attenuation byatmospheric gases,” Tech. Rep., 2017, accessed: 01-04-2019.[Online]. Available: https://www.itu.int/dms_pubrec/itu-r/rec/p/R-REC-P.676-11-201609-I!!PDF-E.pdf

[185] S. Paine, “The am atmospheric model,” March 2018. [Online].Available: https://doi.org/10.5281/zenodo.1193771

[186] E. K. Smith, “Centimeter and millimeter wave attenuation and bright-ness temperature due to atmospheric oxygen and water vapor,” RadioScience, vol. 17, no. 6, pp. 1455–1464, 1982.

[187] ITU, “Recommendation ITU-R P.372-13: Radio noise,” Tech. Rep.,2016, accessed: 01-04-2019. [Online]. Available: https://www.itu.int/dms_pubrec/itu-r/rec/p/R-REC-P.372-13-201609-I!!PDF-E.pdf

[188] F. Box, “Utilization of atmospheric transmission losses for interference-resistant communications,” IEEE Transactions on Communications,vol. 34, no. 10, pp. 1009–1015, October 1986.

[189] P. Boronin, V. Petrov, D. Moltchanov, Y. Koucheryavy, and J. M. Jornet,“Capacity and throughput analysis of nanoscale machine communicationthrough transparency windows in the terahertz band,” Nano Communica-tion Networks, vol. 5, no. 3, pp. 72 – 82, 2014.

[190] H. W. Elayan, C. Stefanini, R. M. Shubair, and J. M. Jornet, “End-to-end noise model for intra-body terahertz nanoscale communication,”IEEE Transactions on NanoBioscience, vol. 17, pp. 464–473, 2018.

[191] E. R. Brown, “Fundamentals of terrestrial millimeter-wave and Tera-hertz remote sensing,” International Journal of High Speed Electronicsand Systems, vol. 13, no. 04, pp. 995–1097, 2003.

[192] P. Boronin, D. Moltchanov, and Y. Koucheryavy, “A molecular noisemodel for Terahertz channels,” in IEEE International Conference onCommunications (ICC), June 2015, pp. 1286–1291.

[193] B. Carli, “Basics of radiative transfer.” [Online]. Available: https://earth.esa.int/dragon/D2_L2_Carli.pdf

[194] L. Ippolito, R. Kaul, R. Wallace, U. S. N. Aeronautics, S. A. Scientific,T. I. Branch, and O. R. Incorporated, Propagation Effects Handbook forSatellite Systems Design: A Summary of Propagation Impairments on 10to 100 GHz Satellite Links with Techniques for System Design, ser. NASAreference publication. National Aeronautics and Space Administration,Scientific and Technical Information Branch, 1985.

[195] A. Straiton, “The absorption and reradiation of radio waves by oxygenand water vapor in the atmosphere,” IEEE Transactions on Antennas andPropagation, vol. 23, no. 4, pp. 595–597, July 1975.

[196] R. B. L. Richard Feynman, Matthew Sands, The Feynman Lectures onPhysics. Ingram Publisher Services US, 2015, vol. 1.

[197] J. Kokkoniemi, J. LehtomaÂlki, and M. Juntti, “Frequency domainscattering loss in Terahertz band,” in Global Symposium on Millimeter-Waves (GSMM), May 2015, pp. 1–3.

[198] F. M. Wu, B. F. Wu, X. Y. Zhao, Z. H. Zhang, H. Zhang, T. Y.Zhang, and Y. Fang, “Analysis on scattering and relationship with granularsize in Terahertz spectra,” in 40th International Conference on Infrared,Millimeter, and Terahertz waves (IRMMW-THz), August 2015, pp. 1–2.

[199] Y. Wang, F. Zhang, Z. Dong, and H. Sun, “Effects of nonsphericity onattenuation characteristics of terahertz atmospheric propagation,” in 41stInternational Conference on Infrared, Millimeter, and Terahertz waves(IRMMW-THz), September 2016, pp. 1–2.

[200] S. Kim and A. ZajiÄG, “Statistical modeling of Terahertz scatterchannels,” in 9th European Conference on Antennas and Propagation(EuCAP), April 2015, pp. 1–5.

[201] S. Ju, S. H. A. Shah, M. A. Javed, J. Li, G. Palteru, J. Robin,Y. Xing, O. Kanhere, and T. S. Rappaport, “Scattering mechanisms andmodeling for terahertz wireless communications,” ICC 2019 - 2019 IEEEInternational Conference on Communications (ICC), pp. 1–7, 2019.

[202] J. Ma, R. Shrestha, L. Moeller, and D. M. Mittleman, “Invited article:Channel performance for indoor and outdoor terahertz wireless links,”APL Photonics, vol. 3, no. 5, p. 051601, 2018.

[203] Z. Xiao, Q. Yang, J. Huang, Z. Huang, W. Zhou, Y. Gao, R. Shu,and Z. He, “Terahertz communication windows and their point-to-pointtransmission verification,” Appl. Opt., vol. 57, no. 27, pp. 7673–7680,September 2018.

[204] A. Moldovan, M. A. Ruder, I. F. Akyildiz, and W. H. Gerstacker, “Losand nlos channel modeling for Terahertz wireless communication withscattered rays,” IEEE Globecom Workshops (GC Wkshps), pp. 388–392,2014.

[205] C. Han and I. F. Akyildiz, “Distance-aware multi-carrier (DAMC)modulation in terahertz band communication,” in 2014 IEEE InternationalConference on Communications (ICC), June 2014, pp. 5461–5467.

[206] C. Liaskos, A. Tsioliaridou, A. Pitsillides, N. Kantartzis, A. Lalas,X. A. Dimitropoulos, S. Ioannidis, M. Kafesaki, and C. M. Soukoulis,“Building software defined materials with nanonetworks,” 2014.

[207] H. Mabed and J. Bourgeois, “A flexible medium access control protocolfor dense terahertz nanonetworks,” in Proceedings of the 5th ACMInternational Conference on Nanoscale Computing and Communication,ser. NANOCOM ’18. New York, NY, USA: ACM, 2018, pp. 1–7.

[208] N. Zhu and R. Ziolkowski, “Photoconductive thz antenna designs withhigh radiation efficiency, high directivity, and high aperture efficiency,”IEEE Transactions on Terahertz Science and Technology, vol. 3, no. 6,pp. 721–730, November 2013.

[209] W. Na, L. Park, and S. Cho, “Deafness-aware MAC protocol fordirectional antennas in wireless ad hoc networks,” Ad Hoc Networks,vol. 24, pp. 121 – 134, 2015.

[210] C. Han, J. M. Jornet, and I. Akyildiz, “Ultra-massive MIMO chan-nel modeling for graphene-enabled Terahertz-band communications,” inIEEE 87th Vehicular Technology Conference (VTC Spring), June 2018,pp. 1–5.

[211] S. Dhillon, M. Vitiello, E. Linfield, A. Davies, M. Hoffmann,J. Booske, C. Paoloni, M. Gensch, P. Weightman, G. Williams, E. Castro-Camus, D. Cumming, F. Simoens, I. Escorcia-Carranza, J. Grant,S. Lucyszyn, M. Kuwata-Gonokami, K. Konishi, M. Koch, C. Schmutten-maer, T. Cocker, R. Huber, A. Markelz, Z. Taylor, V. Wallace, J. Zeitler,J. Sibik, T. Korter, B. Ellison, S. Rea, P. Goldsmith, K. Cooper, R. Ap-pleby, D. Pardo, P. Huggard, V. Krozer, H. Shams, M. Fice, C. Renaud,A. Seeds, A. Stohr, M. Naftaly, N. Ridler, R. Clarke, J. Cunningham,and M. Johnston, “The 2017 Terahertz science and technology roadmap,”Journal of Physics D: Applied Physics, vol. 50, no. 4, January 2017.

46

[212] J. M. Jornet and A. Cabellos, “On the feeding mechanisms forgraphene-based Terahertz plasmonic nano-antennas,” in IEEE 15th Inter-national Conference on Nanotechnology (IEEE-NANO), July 2015, pp.168–171.

[213] M. Esquius-Morote, J. S. GÃsmez-DÄsÂt’az, and J. Perruisseau-Carrier, “Sinusoidally modulated graphene leaky-wave antenna for elec-tronic beamscanning at Terahertz,” IEEE Transactions on TerahertzScience and Technology, vol. 4, no. 1, pp. 116–122, January 2014.

[214] J. M. Jornet and I. F. Akyildiz, “Graphene-based nano-antennas forelectromagnetic nanocommunications in the terahertz band,” in Proceed-ings of the Fourth European Conference on Antennas and Propagation,April 2010, pp. 1–5.

[215] H. Elayan, R. M. Shubair, and A. Kiourti, “On graphene-basedTerahertz plasmonic nano-antennas,” in 16th Mediterranean MicrowaveSymposium (MMS), November 2016, pp. 1–3.

[216] J. M. Jornet and I. F. Akyildiz, “Fundamentals of electromagneticnanonetworks in the Terahertz band,” Found. Trends Netw., vol. 7, no.2-3, pp. 77–233, Dec. 2013.

[217] M. Naftaly, “Device characterization for Terahertz wireless links,” in9th International Congress on Ultra Modern Telecommunications andControl Systems and Workshops (ICUMT), November 2017, pp. 364–369.

[218] B. Zhang, J. M. Jornet, I. F. Akyildiz, and Z. P. Wu, “Mutual couplingreduction for ultra-dense multi-band plasmonic nano-antenna arrays usinggraphene-based frequency selective surface,” IEEE Access, vol. 7, pp.33 214–33 225, 2019.

[219] H. Visser, Array and phased array antenna basics. United States:Wiley, 2005.

[220] V. Petrov, M. Komarov, D. Moltchanov, J. M. Jornet, and Y. Kouch-eryavy, “Interference and SINR in millimeter wave and Terahertz commu-nication systems with blocking and directional antennas,” IEEE Transac-tions on Wireless Communications, vol. 16, no. 3, pp. 1791–1808, March2017.

[221] J. M. Jornet and I. F. Akyildiz, “Information capacity of pulse-basedwireless nanosensor networks,” in 8th Annual IEEE CommunicationsSociety Conference on Sensor, Mesh and Ad Hoc Communications andNetworks, June 2011, pp. 80–88.

[222] ——, “Channel capacity of electromagnetic nanonetworks in the Tera-hertz band,” in IEEE International Conference on Communications, May2010, pp. 1–6.

[223] P. G. Fytampanis, G. V. Tsoulos, G. E. Athanasiadou, and D. A.Zarbouti, “Wireless channel capacity estimation in the Terahertz band,”in International Workshop on Antenna Technology: Small Antennas,Innovative Structures, and Applications (iWAT), March 2017, pp. 339–342.

[224] M. A. Akkas, “Terahertz wireless data communication,” WirelessNetworks, June 2017. [Online]. Available: https://doi.org/10.1007/s11276-017-1548-4

[225] K. Liu, S. Jia, S. Wang, X. Pang, W. Li, S. Zheng, H. Chi, X. Jin,X. Zhang, and X. Yu, “100 gbit/s Terahertz photonic wireless transmissionin the 350-ghz band with extended reach,” IEEE Photonics TechnologyLetters, vol. 30, no. 11, pp. 1064–1067, June 2018.

[226] X. Pang, S. Jia, O. Ozolins, X. Yu, H. Hu, L. Marcon, P. Guan, F. D.Ros, S. Popov, G. Jacobsen, M. Galili, T. Morioka, D. Zibar, and L. K.Oxenlwe, “Single channel 106 gbit/s 16QAM wireless transmission in the0.4 Terahertz band,” in Optical Fiber Communications Conference andExhibition (OFC), March 2017, pp. 1–3.

[227] K. KrishneGowda, A. Wolf, R. Kraemer, J. C. Scheytt, and I. Kallfass,“Wireless 100 gb/s: PHY layer overview and challenges in the Terahertzfreqency band,” in WAMICON, June 2014, pp. 1–4.

[228] J. M. Jornet and I. F. Akyildiz, “Femtosecond-long pulse-basedmodulation for terahertz band communication in nanonetworks,” IEEETransactions on Communications, vol. 62, no. 5, pp. 1742–1754, May2014.

[229] A. K. Vavouris, F. D. Dervisi, V. K. Papanikolaou, and G. K. Kara-giannidis, “An energy efficient modulation scheme for body-centric nano-communications in the Terahertz band,” in 7th International Conferenceon Modern Circuits and Systems Technologies (MOCAST), May 2018,pp. 1–4.

[230] H. Sarieddeen, M.-S. Alouini, and T. Y. Al-Naffouri, “Terahertz-band ultra-massive spatial modulation mimo,” arXiv e-prints, p.arXiv:1905.04732, May 2019.

[231] C. Wang, J. Yu, X. Li, P. Gou, and W. Zhou, “Fiber-thz-fiber link forthz signal transmission,” IEEE Photonics Journal, vol. 10, no. 2, pp. 1–6,April 2018.

[232] S. G. Muttlak, O. S. Abdulwahid, J. Sexton, M. J. Kelly, and M. Mis-sous, “Ingaas/alas resonant tunneling diodes for Terahertz applications:

An experimental investigation,” IEEE Journal of the Electron DevicesSociety, vol. 6, pp. 254–262, 2018.

[233] K. Guan, D. He, A. Hrovat, B. Ai, Z. Zhong, and T. KÃijrner,“Challenges and chances for smart rail mobility at mmwave and thz bandsfrom the channels viewpoint,” in 2017 15th International Conference onITS Telecommunications (ITST), May 2017, pp. 1–5.

[234] M. El-Absi, A. A. Abbas, A. Abuelhaija, K. Solbach, and T. Kaiser,“Distance and tag aware localization in indoor terahertz systems,” in 2018First International Workshop on Mobile Terahertz Systems (IWMTS), July2018, pp. 1–5.

[235] J. M. Jornet and E. Einarsson, “Link and network layers design forultra-high- speed terahertz-band communications networks state univer-sity of new york (suny) at buffalo,” 2016.

[236] J. M. Jornet, “Low-weight error-prevention codes for electromagneticnanonetworks in the terahertz band,” Nano Communication Networks,vol. 5, no. 1, pp. 35 – 44, 2014.

[237] J. M. Jornet and I. F. Akyildiz, “Joint energy harvesting and com-munication analysis for perpetual wireless nanosensor networks in theterahertz band,” IEEE Transactions on Nanotechnology, vol. 11, no. 3,pp. 570–580, May 2012.

[238] J. Xu, J. Jiang, Z. Wang, and Y. Zhao, “Energy harvesting multi-pathrouting for wireless multimedia nanosensor networks in terahertz band,”in 2018 14th International Wireless Communications Mobile ComputingConference (IWCMC), June 2018, pp. 1011–1017.

[239] S. Verma, S. Kaur, G. Dhiman, and A. Kaur, “Design of a novelenergy efficient routing framework for wireless nanosensor networks,”in 2018 First International Conference on Secure Cyber Computing andCommunication (ICSCCC), Dec 2018, pp. 532–536.

[240] A. Moldovan, P. Karunakaran, I. F. Akyildiz, and W. H. Gerstacker,“Coverage and achievable rate analysis for indoor terahertz wirelessnetworks,” in IEEE International Conference on Communications (ICC),May 2017, pp. 1–7.

[241] J. Xu, J. Kan, and Y. Zhang, “Centralized energy harvesting-based tdmaprotocol for terahertz nanosensor networks,” Sensors, vol. 19, no. 20,2019. [Online]. Available: https://www.mdpi.com/1424-8220/19/20/4508

[242] W.-L. Wang, C.-C. Wang, and X.-W. Yao, “Slot self-allocation basedmac protocol for energy harvesting nano-networks,” Sensors, vol. 19,no. 21, 2019. [Online]. Available: https://www.mdpi.com/1424-8220/19/21/4646

[243] C. N. Barati, S. A. Hosseini, S. Rangan, P. Liu, T. Korakis, S. S.Panwar, and T. S. Rappaport, “Directional cell discovery in millimeterwave cellular networks,” IEEE Transactions on Wireless Communications,vol. 14, no. 12, pp. 6664–6678, December 2015.

[244] Q. Xia and J. M. Jornet, “Expedited neighbor discovery in directionalterahertz communication networks enhanced by antenna side-lobe infor-mation,” IEEE Transactions on Vehicular Technology, vol. 68, no. 8, pp.7804–7814, Aug 2019.

[245] S. Balasubramaniam and J. Kangasharju, “Realizing the internet ofnano things: Challenges, solutions, and applications,” Computer, vol. 46,no. 2, pp. 62–68, Feb 2013.

[246] “IEEE standard for low-rate wireless networks,” IEEE Std 802.15.4-2015 (Revision of IEEE Std 802.15.4-2011), pp. 1–709, April 2016.

[247] S. Singh, R. Mudumbai, and U. Madhow, “Interference analysisfor highly directional 60-GHz mesh networks: The case for rethink-ing Medium Access Control,” IEEE/ACM Transactions on Networking,vol. 19, no. 5, pp. 1513–1527, October 2011.

[248] R. Bonjour, M. Singleton, S. A. Gebrewold, Y. Salamin, F. C. Abrecht,B. Baeuerle, A. Josten, P. Leuchtmann, C. Hafner, and J. Leuthold,“Ultra-fast millimeter wave beam steering,” IEEE Journal of QuantumElectronics, vol. 52, no. 1, pp. 1–8, Jan 2016.

[249] S. Rey, “Terapan: Ultra-high data rate transmission with steerableantennas at 300 ghz,” in IEEE 802.15-15-0167-02-0thz, March 2015.

[250] W. Withayachumnankul, “Terahertz metasurfaces for beamforming andpolarization conversion,” in 2018 Asia-Pacific Microwave Conference(APMC), Nov 2018, pp. 112–113.

[251] Z. Hossain, C. N. Mollica, J. F. Federici, and J. M. Jornet, “Stochasticinterference modeling and experimental validation for pulse-based tera-hertz communication,” IEEE Transactions on Wireless Communications,vol. 18, no. 8, pp. 4103–4115, Aug 2019.

[252] A. Galal and X. Hesselbach, “Nano-networks communication architec-ture: Modeling and functions,” Nano Communication Networks, vol. 17,pp. 45 – 62, 2018.

[253] C. Han, A. O. Bicen, and I. F. Akyildiz, “Multi-wideband waveformdesign for distance-adaptive wireless communications in the terahertzband,” IEEE Transactions on Signal Processing, vol. 64, no. 4, pp. 910–922, Feb 2016.

47

[254] M. S. I. Mahfuz and S. Saha, “Channel sharing based medium accesscontrol protocol for wireless nano sensing network,” Global Journal ofComputer Science and Technology, vol. 15, no. 8, 2015.

[255] I. Demirkol, C. Ersoy, and F. Alagoz, “MAC protocols for wirelesssensor networks: a survey,” IEEE Communications Magazine, vol. 44,no. 4, pp. 115–121, April 2006.

[256] M. Kolano, O. Boidol, S. Weber, D. Molter, and G. V. Freymann,“Single-laser polarization-controlled optical sampling system for Tera-hertz - TDS,” in 43rd International Conference on Infrared, Millimeter,and Terahertz Waves (IRMMW-THz), September 2018, pp. 1–3.

[257] S. Priebe, M. Jacob, and T. KÃijrner, “Polarization investigation ofrough surface scattering for thz propagation modeling,” in Proceedingsof the 5th European Conference on Antennas and Propagation (EUCAP),April 2011, pp. 24–28.

[258] M. S. Rabbani and H. Ghafouri-Shiraz, “Improvement of microstripantenna’s gain, bandwidth and fabrication tolerance at Terahertz frequencybands,” in Wideband and Multi-Band Antennas and Arrays for Civil,Security Military Applications, December 2015, pp. 1–3.

[259] P. Lu, V. Rymanov, S. DÃijlme, B. Sievert, A. Rennings, and A. StÃuhr,“Terahertz beam forming and beam switching using lens-assisted quasi-optical thz transmitter,” in International Topical Meeting on MicrowavePhotonics (MWP), October 2017, pp. 1–4.

[260] S. Nie, J. M. Jornet, and I. F. Akyildiz, “Intelligent environmentsbased on ultra-massive mimo platforms for wireless communication inmillimeter wave and terahertz bands,” in IEEE International Conferenceon Acoustics, Speech and Signal Processing (ICASSP), May 2019, pp.7849–7853.

[261] Y. Zhao, B. Ai, D. Fei, Y. Liu, and N. Li, “Adaptive beamforming basedon subband structure in smart antennas,” in 32nd General Assembly andScientific Symposium of the International Union of Radio Science (URSIGASS), August 2017, pp. 1–2.

[262] J. Ma, R. Shrestha, W. Zhang, L. Moeller, and D. M. Mittleman,“Terahertz wireless links using diffuse scattering from rough surfaces,”IEEE Transactions on Terahertz Science and Technology, vol. 9, no. 5,pp. 463–470, Sep. 2019.

[263] Z. Rong, M. S. Leeson, and M. D. Higgins, “Relay-assisted nanoscalecommunication in the Terahertz band,” Micro Nano Letters, vol. 12, no. 6,pp. 373–376, 2017.

[264] M. Park and H. K. Pan, “Effect of device mobility and phased arrayantennas on 60 ghz wireless networks,” in Proceedings of the 2010 ACMInternational Workshop on mmWave Communications: From Circuits toNetworks, ser. mmCom ’10. New York, NY, USA: ACM, 2010, pp.51–56.