Comparison Between L TE and WiMAX

28th NATIONAL RADIO SCIENCE CONFERENCE (NRSC 2011)

April 26-28, 2011, National Telecommunication Institute, Egypt

Comparison between L TE and WiMAX based on System Level Simulation

Using OPNET modeler (release 16)

Eng. Mohammed Torad, Dr. Ahmed EI Qassas, Prof Dr. Hadia Al Henawi Ain Shams University, Department of Electronic and Communication Engineering, Cairo, Egypt

Al Baha Private College of Science, Department of Computer Engineering, Al Baha, KSA Ain Shams University, Department of Electronic and Communication Engineering, Cairo, Egypt

ABSTRACT

In this paper a comparative study between L TE and WiMAX, the two leading wireless broadband technologies, is introduced. The performance metrics used for the evaluation are the response time and the throughput derived from OPNET modeler release 16 system level simulator. To perform the comparison 4 scenarios were developed the fIrst two are for 7 cell LTE and 7cell WiMAX networks while the second two are for 19 cell L TE and 19cell WiMAX networks. The numerical results are obtained for links between mobiles at fIxed relative positions in the four scenarios. Through this analysis we found that LTE networks gave shorter response time than WiMAX. The throughput is also compared and results showed that WiMAX outperformed L TE in this respect.

Keywords: L TE, WiMAX, system level simulation, OPNET, response time, and throughput.

I. INTRODUCTION

In recent few years telecommunication authorities are busy deciding how to emerge to 4G environment motivated by the exponential increase in the demand for advanced telecommunication services which require wider spectrum and higher quality of service. On the other hand, telecommunication industry experts are trying hard to standardize new mobile wireless systems that can cope with the desires and ambitions of telecommunication users and pave the way for evolving new technologies. New applications required to be supported by new mobile systems include a variety; VoIP, video conference, multimedia messaging, multiplayer games, virtual private networks (VPN), etc. [9] All these application require higher throughput, wider BW, smaller delay and innovative transmission methods that will give higher spectral efficiency and good quality. Two leading emerging technologies are: L TE (Long Term Evolution standardized by third generation partnership project (3GPP)) and WiMAX (the IEEE802.16e, the worldwide interoperability for microwave access) are considered able to fulfIl the 4G requirements announced by ITU-R which is known as international mobile telecommunications advanced (IMTS) [9]. During the standardization process of both technologies, computer simulations are necessary to test the validity of proposed algorithms and procedures and to help improving them. [1], [2], [8], [10], [12] Two classes of simulator are built and implemented. The fIrst is the link level simulation that is used to investigate physical layer issues such as multiple input multiple output (MIMO), adaptive modulation and coding (AM C), channel estimation and equalization, and channel modelling, [10], [12]. The second is the system level simulation which investigates networking performance such as scheduling, call admission control, mobility handling, interference management, and power control. [1], [2], [8] This separation into link level and system level simulation has a big benefIt as it reduces the computation complexity that will increase exponentially if the two types of simulation were merged together. As an example modeling of the fast fading behaviour of the wireless channel for each individual user requires high computational power and time. This computational effort is mandatory in link level simulation issues but if it is added in the system level the computational effort will be excessive [2] Therefore abstraction is used in the simulation in the sense that some issues are modeled and simulated off line and the results are included in the system level simulation run, as an assumption or constant values or it is not included in the simulation at all. This method will lead to system level simulation with valid performance results and low computational complexity [2]

Our main job in this paper is to perform system level simulation of the two emerging technologies and compare between them. We use for this task the OPNET modeler releaseI6 [4]. We developed four scenarios the fIrst two are for 7- cell LTE and 7-cell WiMAX networks while the second two are for 19- cell LTE and 19-cell WiMAX networks. The performance metrics that will be considered are: the response time and the throughput. The physical channel air interface effect is abstracted here (we assumed ideal channel or a pipeline air interface) for both systems; LTE and WiMAX.

The rest of this paper is organized as follows: Section 2 will include system description that will summarize the simulation scenarios for both L TE and WiMAX. Section 3 will describe the simulation platform of the two



systems (LTE and WiMAX) using OPNET modeler release 16.Section 4 will include the simulation results and analysis of results. Section 5 will conclude the paper.

II. SYSTEM DESCRIPTION:

11.1. L TE SYSTEM:

The LTE (long term evolution) was first documented and standardized by 3GPP in release 8 in december2008. This document; release 8; encompasses the evolved UTRAN (E-UTRAN) and Evolved Packet Core (EPC) of the L TE networks, [21], [9]. The adaptation of L TE standard to the requirements of (lTU -R) IMTs (4G mobile) is performed through the LTE advanced, 3GPP release 9 and release 10 [22], [23] and [24]. 3GPP release 8 defmes a new physical layer based on orthogonal frequency division multiple access (OFDMA) for the downlink and single carrier frequency division mUltiple access (SC-FDMA) for uplink. L TE design Targets include: mobile access at speeds up to 350KInIh, throughput 100Mb/s in downlink and 50Mb/s in uplink increasing to a maximum of 326.4Mb/s and 86.4Mb/s using MIMO 4x4 in the downlink and uplink respectively using BW=20MHz .[9], [21]. The architecture of the LTE system to be simulated is compatible with 3GPP release 8 and is depicted in fig. 1 below [9], [21].

11.2. WiMAX SYSTEM (IEEE802.16e). [16], [17]

The WiMAX (worldwide interoperability microwave access) is a Metropolitan wireless network based on IEEE 802.16-2004 standard and the IEEE 802. 16e-2005 standard which modify the 2004 standard and adds new features to the system like mobility. These two standards defme the physical and the MAC layer as follows: Physical layer uses SOFDMA, the allocated frequency band is 2-11 GHZ for non line of sight (NLOS) and frequency range 10-66GHZ for line of sight (LOS) [most implemented WiMAX networks work are in the 3.5GHZ range]. It support channel bandwidth from 1.5-20 MHZ. [25], [24]. The wireless coverage distance is up to 8 KIn for LOS and up to 50 KIn for NLOS with bit rates of up to 70 Mb/s for 20 MHZ BW. The MAC layer is designed to support packet transmission and support IP and ATM protocols. The MAC layer structure comprises 3 sub layers; the security sub layer, the common part sub layer and MAC convergence sub layer. The security sub layer performs authentication, encryption and key exchange between subscriber station and base station. The common part sub layer perform bandwidth allocation, connection establishment and connection maintenance. The convergence sub layer transforms data, assign service class to the appropriate traffic flow with required quality of service. The WiMAX network model [26] consists of the following entities: the subscriber station (SS), the access service network (ASN) which comprises the base station (BS) and the Access Service Network (ASN) gateway and the connectivity service network (CSN) which consists of visited network service provider and home network service provider.Fig.2 shows the WiMAX network reference model [26] [24]

� PDNGW \- \-

MME/scrving W

eNS

Fig.l L TE system architecture as per release 8.

; .................. .......... ��1....... . ...... .. Y'l�l!��J�.�f .. ........ .. H!?!!1!! N5P i ............... ........... -f - ...... ....... ......... : i

� R2

i i 55/ Rl

ME .... .[. ..... .

I

R3

.. ··-1-· ·· .. I ASN

Another ASN

NAP

CSN

RS

.. ··-1-··· .. CSN I

I ASP network OR internet

Fig.2 WIMAX network reference model as per [25]



III. SIMULATION PLATFORM USING OPNET MODELER 16: [4] OPNET modeler version 16 simulator is a packet based event driven dynamic system level simulator which

accurately and efficiently simulate the behaviour of various types of real world networks. OPNET Modeler is a product of the OPNET Technologies Inc. It has a Graphical User Interface (GUI) with a "user friendly" sense. It has an enormous library at the service of the user. For the research to be done in this paper, "OPNET Modeler version 16" was available under license. We will use this simulator to simulate the two systems; LTE and WiMAX with the purpose to compare their performance. Using OPNET modeler 16 we developed four scenarios; two of them are for L TE networks (one for 7 -cell network structure and the other for 19-cell networks structure). The other two scenarios are for WiMAX networks (one for 7-cell network structure and the other for 19-cell networks structure).

111.1. OPNET LTE NETWORK SIMULATION: [4] Fig. 3 shows the proposed L TE 7 -cell structure and fig. 4 shows the proposed L TE 19-cell structure the two

systems are constructed according to fig. 1 above. The system main parts are: the user equipment UE, the access network E-UTRAN which consists of enhanced node B (eNB) and the X2 interface between them, and the evolved packet system (EPC) which comprises the mobility management entity (MME) and the serving gateway (ASN S-GW).

111.1.1. ENTITIES OF THE SCENARIO: [4] User equipment, eNB, EPC, MAC function and physical layer functions.

• User equipment UE: For each user equipment UE EPC bearers are created (up to 8 bearers, one of them is called the default bearer) the type of the EPC bearer can be GBR (Guaranteed Bit Rate) or non-GBR (None Guaranteed Bit Rate). The default bearer is created as soon as the UE gets attached to an EPC through the eNB (this is the entity through which data is transferred). Each bearer has the following attributes: allocation retention priority and quality of service QOS class identifier (corresponding to QCI quality channel identifier) which associates the bearer with a QOS class definition. In this simulation we use QOS identifier corresponding to the voice and live video type of service. The EPC bearers are considered bidirectional. For each bearer one traffic template is associated.

• eNB: Functions of eNB that are simulated are the admission control. When requiring creating a GBR bearer, it goes through the admission control mechanism which allocates node resources for this bearer and deducts them from the available cell resources. While calculating the cell capacity resources all the overheads are taken into consideration.

• EPC: Functions of the evolved packet core EPC that are simulated are: Mobility management entity and session management functions (request for creation and activation of a bearer).

• Radio link control functions: The scheduler performs the segmentation and concatenation procedures using a dynamic PDU size and decides the allocation size.

• Medium access control functions: Mapping the EPC radio bearer into created logical channels and logical channels into transport channels is performed. Also random access procedures are implemented by the UE MAC. In addition the UE MAC can make scheduling request transmission such that resources are allocated.

• Physical layer functions: The following physical layer functionalities are simulated: OFDMA, cyclic prefix length, the air interface parameters are configured (OFDMA default parameters, BW = 20 MHz, SISO, the modulation and coding scheme (MCS) is chosen compatible with the service type). All types of physical channels are simulated [PDSCH, PDCCH, PRACH, PUSCH AND PUCCH]. For this simulation we use frequency division duplexing mode of operation (FDD).

111.2 OPNET WiMAX NETWORK SIMULATOR:

Fig.5 shows the proposed WiMAX 7-cell structure network and fig.6 shows the proposed 19-cell WiMAX network. The two structures are developed according to fig.2 above.

111.2.1. THE MAIN ENTITIES OF THE WiMAX: The main entities of the WiMAX are the subscriber station (SS), base station (BS), the ASN gateway and the

ASN features.

• Base station (SS): can be sectorized or Omni-directional, fixed or mobile (802.16j functionality). In the simulation scenario we will be restricted to fixed Omni-directional base station.

• Subscriber station: it can be subscriber station with server functionality Or SS with WLAN router functionality or SS with workstation functionality.

For each SS-BS connection MAC service class can be configured, OFDMA physical profile, adaptive coding and modulation can be configured. The physical layer parameters can be configured.



To develop a WiMAX network we create the topology, the node mobility (SS), add traffic and then configure WiMAX parameters.

·3000 -2000 -1000 o 1000 2000 3000

L TE A.!,Wbutes:

3000

2000 ProflU. Definition

Applications:

1000

o

·1000

·2000

Fig.3 proposed L TE 7- cell network 000 o 5000

5000

Fig.4 proposed L TE 19- cell network

·3000

3000

2000

1000

·1000

·2000

·!iQOO

5000



·2000 ·1000 1000 2000 3000

Fig.S proposed WiMAX 7 - cell network

4000

5000 10000

Applications Profiles

WiMAX_Config README

Fig.6 proposed WiMAX 19- cell network



111.2.2. WiMAX CONFIGURABLE PARAMETERS:

Service class (Gold, Silver or Bronze), efficiency mode (we choose framing module enabled which comprises MAC service flow with PHY abstracted), physical layer profile (SOFDMA configuration), Definition of service flows between base station and subscriber station, mobility configuration (allows the association of SS with BS) We choose service flow compatible with voice and video streaming. Adaptive modulation and coding (AMC) is enabled for each service flow, it is configured at the SS parameters and it is copied by default to the corresponding base station configuration. The physical layer (channel) parameters like path loss, transmitter power, and multipath channel model can be configured, but for purposes of comparison with L TE network we abstracted the physical channel in both networks. The access service network gateway has advanced feature that can be configured like ARQ, HARQ, power saving mode and 802.11j (support the relay of WiMAX packets (data and control». Physical layer parameters are configured to be similar to those configured for LTE physical layer.

IV. SIMULATION RESULTS AND ANALYSIS:

In this section simulation results will be presented. At first we will define the performance metrics that will be used for the comparison. The response time is the round-trip time taken for a response to get to the

destination and return to the source. The second parameter is the throughput, for simulations done in OPNET

Modeler the difference between the traffic sent from a source and traffic received at the intended destination is known as TRAFFIC DROPPED and could be due to mainly the fading wireless channel and also due to different network effects such as congestion, interference etc, which could occur at different layers. The throughput is traffic sent minus the dropped traffic divided by traffic sent. Fig.7 shows the response time (The probability density function) for a sample of a set of links starting at workstation 3_2 during the simulation period for WiMAX 19-cell network. The average values for these curves are depicted in table 1. The same table lists the results obtained for the four scenarios (L TE 7 -cell, L TE 19-cell, WiMAX 7-cell and WiMAX 19-cell).The results show that LTE gave shorter response time than WiMAX. These results may be due the scheduling algorithms adopted by WiMAX standards. L TE readings do not exceed lOOms which satisfy the requirements of IMTs regarding the delay budget and the requirements stated in release 8 regarding the maximum Latency [21], [9]. The readings for WiMAX response time range from50 ms to several seconds which might affect real time applications. Response time can mean various things. If we are using an FTP application, response will usually consist of just one direction. In contrast, for a VoIP, voice or video conference type application, response will match more closely to the round trip time. There are many factors in play, such as ARQ/HARQ retransmissions, TCP retransmissions etc, that can make all data to be received correctly albeit with higher delay.

so 1111

Object Mobile _10_1

(MobiIUO_l<-->Mobile_3_2)

ISO 200 250 3SO 400 4SO S50

Object Mobile _16_1

(Mobile _16 _1 <->Mobile_3

600 6SO

Fig.7 Simulation results response time (probability density function PDF) For WIMAX 19 - cell scenario (3D)



It is worth mentioning here that the above mentioned readings come from random processes performed by the simulator. These readings may vary for each simulation run but we depicted the most probable values obtained from many simulation runs. The two standards for WiMAX; IEEE802.16-2004 and IEEE802.16e focused on the physical and MAC layers performance only while the new IEEE802.l 6m TG is focusing on the end- to - end performance improvement which includes the scope of the network and application indeed [17]. Many papers discussed how to improve the delay performance of WiMAX. The improvements include using new scheduling algorithms [17], using TCP acknowledge manager [16] or improving the service flow reservation! modification process by creating one auxiliary service flow in the WiMAX network [30] and others. These improvements deal with the real reasons of long response time of WiMAX

Table 1 simulation results for the four scenarios (response time)

Source Node

VE 3 2 VE 3 2 VE 3 2 VE 3 2 VE 3 2 VE 3 2 - -VE 3 2 - -VE 3 2 - -VE 3 2 - -

Destination Node

VE 2 5

VE 4 3 VE 5 1

VE 6 4 VE 7 1

VE 8 1 - -VE 10 1 - -VE 14 1

VE 16 1 - -

LTE 7-cell Response time (m seconds)

77 72 42

88

53 -

-

-

-

LTE 19-cell WiMAX WiMAX

Response 7-cell 19-cell

time Response Response

(m seconds) time time (m seconds) (m seconds)

60 12300 3406 63 3090 2060 39 60 90

89 570 3600 75 3010 90 38 - 1730

61 - 90

75 - 2243

43 - 5566

Regarding the reuse factor (7/19) cell, we cannot judge from the response time readings depicted in table 1 if the response time (delay) increases or decreases as the size of the network increases because the value of the reading depends not only on the size of the network but also on the instantaneous downlink / uplink available resources at the BS in terms of bandwidth (packets/second) at the interface between BS and ASN -GW at the time of sending or receiving the traffic [16]. Regarding the throughput results, Fig.8 shows the throughput for a sample of a set of links starting at workstation 3_2 during the simulation period for L TE 19-cell network. The average values for these curves are de icted in table 2.

bject UE_S_l bject UE_6_ 4

bject UE_14_1 bject UE_16_1

UE_12_1

Fig.8 Simulation results throughput for LTE 19 - cell scenario (3D)

Table 2 summarizes the results obtained from the simulation for the four scenarios. The results show that WiMAX has better results regarding throughput. It is worth mentioning here that an abstraction is made in this simulation. It considered in WiMAX simulation that there is no dropped traffic due to network effects like



congestion and negative acknowledgements. The results for simulated dropped traffic are zero all time which mean enhancement for the throughput result. For verification of these results we found that in [29] a similar

Table 2 simulation results for the four scenarios (throughput)

Node

UE 2 5

UE 3 2

UE 4 3 UE 5 1

UE 6 4 UE 7 1

UE 8 1 - -

UE 10 1 - -

UE 14 1

UE 16 1 - -

LTE 7 cell Throughput

(packets/sec) 9.717 9.903

12.237

6.44 16.982 3.083

-

-

-

-

LTE 19 cell WiMAX 7 cell WiMAX 19 cell Throughput Throughput Throughput

(packets/sec) (packets/sec) (packets/sec) 6.401 43.667 41.553

17.839 42.425 42.352

6.031 41.557 41.523

2.65 43.380 41.787

4.316 42.607 42.165

2.545 43.818 42.18

7.571 - 42.138

6.141 - 41.987

4.89 - 42.11

5.184 - 42.595

comparison was made between UMTSIHSPA+ and WiMAX IEEE802.16e in throughput and coverage and the results were in favour of UMTS/HSPA+.

v. CONCLUSIONS

In this paper we presented a comparison between L TE and WiMAX networks; the two leading wireless broadband technologies. The metrics used for the comparison are response time and the throughput. We presented the standard models of the two systems in compliance with their respective standards, and then we discussed the methodologies adopted by OPNET modeler while building the simulation models of the two competing systems. Simulation results using OPNET modeler were presented; they showed that LTE always gave better performance regarding the delay or response time. We noted that this is a problem for WiMAX because its standards focused on the performance of the physical and MAC layers leaving aside the networking and application aspects as it without customization. We noted, also, that a large number of research papers dealt with this aspect. Regarding the throughput results, we showed that WiMAX outperformed L TE in this respect On the other hand, we noted that an earlier paper compared between UMTSIHSP A + and W iMAX IEEE802.16e in throughput and coverage and the results were in favour of UMTSIHSPA+.

REFERENCES

[1] Xing Zhang, Rinshui Zhu, Shiming Liu, Wenbo Wang, "system level performance analysis of OFDMA for UTRAN systems", proceedings of the 6th WSEAs international conference on multimedia systems a n d s i g n a l p r o c e s s i n g , H a n g z h o u , C h i n a , A p r i l 1 6 - 1 8 , 2 0 0 6 , p pI9 2 - 1 9 7 .

[2] Josep Colom ikuno, Martin Wrulich, Markus Rupp"system level simulation of LTE networks", IEEE vehicular technology conference VTC2010 spring, (3F: Scheduling, Chair: Ying Wang, Beijing University of Posts and Telecommunications), Taipi, Taiwan, May 2010.

[3] Jan Ellenbeck, Johannes Schmidt, Ulrike Korger and Christian Hartmann,"A concept for efficient system level simulation of OFDMA system with proportional fair fast scheduling" institute of communication networks, Munich, Germany, 2009.

[4] http://www.opnet.com (opnet modeler release 16) [5] Tejas Bhandare "LTE and WiMAX comparison ", master thesis at Santa Clara university, December

2008

[6] A.omri, R.Bouallogue, R.hamilaoma, M.Hasan"channel estimation for LTE uplink system by perception neural networks", international journal of wireless and mobile networks vol.2 No.3 august 2010.

[7] Wael Elsayed Othman, Tarek abd Rahman, "optimization of guard time length for mobile WiMAX system over multipath channel" proceeding of IMECS, Hong Kong, 2008

[8] "LTE system level simulator documentation v1.0r295"a documentation on how to use the Long Term Evolution (LTE) System Level simulator [Online]. Available: at http://www.nttuwien.ac.atlltesimulator/ as well as some insight on its structure and the assumptions that were made while developing it, Institute



of communication and radio frequency engineering, Vienna University of technology Austria. [9] David Martin Sacristan, Jose F. Monserrat, Jorge Cabrejas, Daniel Calabuig, Salvador Gorigas and

Narcis Cardona,"on the way towards fourth generation mobile: 3 GPPL TE and L TE advanced" EURASIP journal on wireless communications and networking, volume 2009.

[10] Albert Serra pages, "A long term evolution link level simulator" master thesis from university Politecnica de Catalunya, Feb 2009.

[11] Ramjee Parsad, Fernando D. Velez, "WiMAX Networks "techno economic vision and challenges, Springer, 2010.

[12] Christian Mehlfuhrer, Martin Wrulich, Joseph Colon Ikuno, Dagmar Bosanska, and Markus Rupp, "simulating the long term evolution physical layer" 17th European signal processing conference EUSIPCO August 2009, pages 1471-1478

[13] Moray Rumney," LTE and the evolution to 4G wireless" John Wiley & sons Ltd., July 2009. [14] Sock-Yee, Peter Muller, Hamid R. Sharif, "WiMAX Security and Quality of service" an end - to - end

perspective, Wily, 2010. [15] WIMAX forum,"mobile WIMAX - part I: technical overview and performance evaluation" WIMAX

forum publications, August 2006. [16] Bong Ho Kim, Jungman Yun and Yerang Hur, "capacity estimation and TCP performance enhancement

over mobile WiMAX networks" IEEE communication magazine, June 2009, page 131-142. [17] Shaswar Baban, "Design and implementation of scheduling algorithms for the IEEE802.16e (mobile

WIMAX) networks" master thesis in the department of electronic systems at the university of west minister, September 2008.

[18] Raj Jain, "WiMAX system level modeling methodology: A tutorial" WiMAX forum publication October 2007.

[19] JinyunZhang, "Fast handover protocols for WiMAX networks" patent from Cambridge, MA US, August 2010

[20] Seok Jun Lee, Manish Yuming Zhu, Jing Fei Ren and Yang Sun, "Forward error correction decoding for WiMAX and 3 GPP LTE modems " R&D center Texas instruments, Rice university 2009.

[21] 3GPP TR 25.913, "Requirements for Evolved UTRA (EUTRA) and Evolved UTRAN (E-UTRAN)," v.8.0.0, December 2008.

[22] 3GPP TR 36.913, "Requirements for Further Advancements for E-UTRA," v.8.0.1, March 2009 [23] 3GPP TS 36.213, "Physical layer procedures," v.8.6.0, March2009. [24] 3GPP TS 36.211, "Long Term Evolution physical layer; General description," v.8.6.0, March 2009. [25] IEEE STD 802. 16-2004"IEEE standard for local and metropolitan area networks part 16: air interface for

fixed broadband Wireless access system", IEEE standard 802.16-2004, October 2004. [26] IEEE STD 802. 16-2005"IEEE standard for local and metropolitan area networks part 16: air interface for

Fixed broadband Wireless access system- physical and medium access control layer for combined mobile and Fixed Operation in licensed bands", IEEE standard 802.16e-2005, Feb 2006.

[27] Carlos G. Bilich, "TCP over WiMAX networks", university of Trento, Italy2004 [28] Z.Becvar, I.Zelenka and R.Bestak," comparison of handover in UMTS and WiMAX" Czech Technical

university in Prague, 2006. [29] R.Preguica, L.M.Correia and C.Caseiro, "comparison between UMTS/HSPA+ and

WiMAXlIEEE802.16e in mobility scenarios", institute of telecommunication, Lisboa, Portugal, 2008. [30] Miguel Almeda, Pedro Neves, Daniel Carujo and Sussana Sargento, "Improving the experience of real

time services in WiMAX networks" institute of telecommunications-polo avero, university of Santiago, Portugal, 2006.

Documents

Comparison Between L TE and WiMAX