Ericsson Review: special 90th anniversary edition

The communications technology journal since 1924

Non-line-of-sight microwave backhaul for small cells 12

Software-defined networking: the service provider perspective 20

HSPA evolution: for future mobile-broadband needs 26

Next generation OSS/BSS architecture 38

Carrier Wi-Fi: the next generation 48

Nine decades of innovation 11, 19 & 47

Next generation video compression 33

Delivering content with LTE Broadcast 4

901924-2014

“The object of this magazine is to spread information concerning the work and activities of this and associated enterprises, and to furnish a connecting link between these latter and the head firm.”

These lines are taken from the introduction to The L. M. Ericsson Review – Tidskrift för Allmänna Telefonaktiebolaget L. M. Ericsson – when the first issue of the new journal was published in 1924.

In his historical article to celebrate this journal’s 50th anniversary (1974), Sigvard Eklund, former editor of Ericsson Review (1943-1972), wrote the following words:

“Apart from an article on ‘the devel-opment and present size of the LM Ericsson Group,’ there was a 10-page description of the company’s auto-matic 500-line selector system, illus-trated by a few photographs of the recently opened automatic exchange in Rotterdam, one of the first major exchange equipments to be delivered up to that time.”

In the 40 years since then – and the 90 years since this journal first started promoting technology – the world we live in has been transformed by tech-nology to such a degree that I some-times find it difficult to recognize the old one. I would like nothing more than to be able to give you a glimpse of what we will be writing about 90 years from now… what will the 22nd century bring? What generation of technology will have been reached by then, what business models will we use, how will we pay for things, and what sort of devices will we connect with? These are just some of my ques-tions, and I wonder even if they will be relevant; maybe we won’t even use devices, as connectivity will simply exist in everything. Even if I don’t have an open window on the next centu-ry, the research and development we carry out at Ericsson today is aimed at the next generation, which promises

Ulf EwaldssonChief Technology Officer

Head of Group Function Technology at Ericsson

Celebrating 90 years of technology insights

to continue along the current path of evolution: to be data driven, video heavy and influenced by the gaming world.

The articles in this edition address a wide range of telecommunication issues, but they all have one thing in common and that is performance. Getting data through the network fast and efficiently so that the best user experience can be delivered to sub-scribers is a recurring theme no mat-ter what part of network architecture is being discussed. From future OSS/BSS architecture to integrated Wi-Fi and to packets stuffed with data, the message is clear… the faster the net-work can serve one subscriber, the faster it can move on to the next.

One thing I am convinced about, however, is convergence. And not just in terms of fixed and mobile, but everywhere. Industries and technolo-gies are merging. The lines between TV, the internet, and telecommuni-cation will not exist for much longer. Education, work and family life are all coming together and the key is individualism.

With a connection, every individ-ual on the planet has the potential to take control over their life. The tradi-tional models of work and education are being challenged. Connectivity is providing individuals with more choices, greater flexibility and the abil-ity to mix things up in a way that suits them, their budget, their lifestyle and their goals. We are not there yet, and there are many pieces that need to be in place, but right now we are lay-ing the foundation for the Networked Society. Mobile subscriptions are set to rise to 9.3 billion and mobile data traf-fic to grow by 45 percent (CAGR) by 2019. The opportunities are becoming available for more people, and connec-tivity is becoming a way of life.

This edition is a celebration of 90 years of technology innovation. I hope you enjoy it.

The most frequent users interact with their smartphone more than 150 times a day, or an average of every seven minutes during the daytime.*

Editorial

*Ericsson Mobility Report, November 2013

E R I C S S O N R E V I E W • 90TH ANNIVERSARY • 2014

The communicat ons techno ogy journal since 1924

Non line of sight microwave backhaul for small cells 12

Software defined networking the service provider perspective 20

HSPA evolution for future mobilebroadband needs 26

Next generation OSS/BSS architecture 38

Carrier Wi Fi the next generation 48

Nine decades of innovation 11 19 & 47

Next generation video compression 33

Delivering content with LTE Broadcast 4

901924-20 4

CONTENTS 90TH ANNIVERSARY 2014

a collection of articles from 2013

4 Delivering content with LTE BroadcastThe data volume in mobile networks is booming – mostly due to the success of smartphones and tablets. LTE Broadcast is one way of providing new and existing services in areas that can at times be device dense, such as stadiums and crowded city centers. Built on LTE technology, LTE Broadcast extends the LTE/EPC with an efficient point-to-multipoint distribution feature that can serve many devices with the same content at the same time. This article was originally published on February 11, 2013.

11, 19 & 47 Nine decades of innovationAutomatic exchanges to smart networks.

12 Non-line-of-sight microwave backhaul for small cellsThe evolution to denser radio-access networks with small cells in cluttered urban environments has introduced new challenges for microwave backhaul. A direct line of sight does not always exist between nodes, and this creates a need for near- and non-line-of-sight (NLOS) microwave backhaul. This article was originally published on February 22, 2013.

20 Software-defined networking: the service provider perspective An architecture based on software-defined networking (SDN) techniques gives operators greater freedom to balance operational and business parameters, such as network resilience, service performance and QoE against opex and capex. With its beginnings in data-center technology, SDN has developed to the point where it can offer significant opportunities to service providers. This article was originally published on February 21, 2013.

26 HSPA evolution for future mobile-broadband needsAs HSPA evolution continues to address the needs of changing user behavior, new techniques develop and become standardized. This article covers some of the more interesting techniques and concepts under study that will provide network operators with the flexibility, capacity and coverage needed to carry voice and data into the future, ensuring HSPA evolution and good user experience. This article was originally published on August 28, 2013.

33 Next generation video compressionRequiring only half the bitrate of its predecessor, the new standard – HEVC or H.265 – will significantly reduce the need for bandwidth and expensive, limited spectrum. HEVC (H.265) will enable the launch of new video services and in particular ultra-HD television (UHDTV). This article was originally published on April 24, 2013.

38 Next generation OSS/BSS architectureWhen two large companies merge, it often takes a while – years in some cases – before processes get redesigned to span all departments, and the new organization settles into a lean and profitable machine. And the same is true of OSS/BSS. These systems have been designed for two different purposes: to keep the network operational and to keep it profitable. But today’s demanding networks need the functions of both of these systems to work together, and to work across the varying life cycles of products and services. This article was originally published on November 25, 2013.

48 Carrier Wi-Fi: the next generationPutting the network in control over whether or not a device should switch to and from Wi-Fi, and when it should switch, will make it easier for operators to provide a harmonized mobile broadband experience and optimize resource utilization in heterogeneous networks. This article was originally published on December 20, 2013.

To bring you the best of Ericsson’s research world, our employees have been writing articles for Ericsson Review – our communications technology journal – since 1924. Today, Ericsson Review articles have a two-to-five year perspective and our objective is to provide you with up-to-date insights on how things are shaping up for the Networked Society.

Address :EricssonSE-164 83 Stockholm, SwedenPhone: +46 8 719 00 00

Publishing:Ericsson Review articles and additional material are pub ished on www ericsson.com/review. Use the RSS feed to stay informed of the latest updates.

Articles are also available on the Ericsson Technology Insights app for Android and Apple

tablets. The ink for your device is on the Ericsson Review website:www.ericsson.com/review. If you are viewing this digitally, you can:download from Google Play ordownload from the App Store

Publisher: U f Ewaldsson

Editorial board: Håkan Andersson, Hans Antvik, Ulrika Bergström, Joakim Cerwall, Deirdre P. Doyle, Dan Fahrman, Anita Frisell, Jonas Högberg, U f Jönsson, Magnus Karlsson, Cenk Kirbas, Sara Kullman, Kristin Lindqvist, Börje Lundwall, Hans Mickelsson, U f Olsson, Patrik Regårdh, Patrik Roséen and Gunnar Thrysin

Editor: Deirdre P. Doyledeirdre.doyle@jgcommunication se

Chief subeditor: Birgitte van den Muyzenberg

Contributors: John Ambrose, Håkan Andersson, Paul Eade, Ian Nicholson, Gunnar Thrysin and Peter Öhman

Art director and layout: Jessica Wiklund and Carola Pilarz

Illustrations: Claes-Göran Andersson

Printer: Edita Bobergs, Stockholm

ISSN: 0014-0171

Volume: 91, 2014


Delivering content with LTE BroadcastEricsson has demonstrated LTE Broadcast with evolved Multimedia Broadcast Multicast Services at a number of international trade shows. These demos have shown the solution’s potential to create new business models for telcos and ensure consistent QoS, even in very densely populated places like sports venues.

respondents stated they would watch more TV if the content was provided on their mobile device, and 61 percent said they would switch operator to gain access to mobile-TV services. The major-ity of respondents said content they would find interesting to watch while on the move includes local news and weather information, movies, national news, sitcoms and sports.

To meet this growing demand for mobile TV, operators are rapidly updat-ing their offerings, continuously add-ing new services and content to live and on-demand streams – increasing the volume of information transport-ed. Naturally, this causes network utili-zation to rise, requiring more efficient ways to deliver content, while network dimensioning becomes all the more cru-cial, and new business models are need-ed to maintain ARPU.

Given the direction in which the industry is clearly moving, Ericsson has developed an end-to-end LTE Broadcast

the broadcast streams within the SFN that are of interest. In this way, devices download only relevant data – not every-thing within the area to then just throw unwanted data away. This ensures that devices work in a battery-efficient way.

Business incentivesThe coextending evolution of mobile technologies and devices has made it possible for people to consume video using handheld equipment without compromising their experience. Based on an Ericsson ConsumerLab study1, the most recent Ericsson Mobility Report2, states that video is the biggest contribu-tor to mobile-traffic volumes, account-ing for more than 50 percent. And the growth of traffic is expected to contin-ue, increasing 12-fold by 2018.

According to another study, carried out by Mobile Content Venture3, more than half of US consumers would con-sider viewing programs on their smart-phones and tablets – 68 percent of

T HOR ST E N LOH M A R, M ICH A E L SL SSI NGA R, V E R A K E N E H A N A N D ST IG PU UST I N E N

BOX A Terms and abbreviations

AL-FEC Application Layer FECAPI application program interfaceARPU average revenue per user BLER Block Error RateBM-SC Broadcast Multicast Service CenterCDN content distribution networkeMBMS evolved MBMSeNB eNodeB EPC Evolved Packet Core EPS Evolved Packet SystemFDD frequency division duplexFEC forward error correctionFIFA Fédération Internationale de Football Association

FLUTE filedeliveryoverunidirectional transport HEVC HighEfficiencyVideoCodingIMB integrated mobile broadcastISD inter-site distanceISI inter-symbol interferenceM2M machine-to-machineMBMS Multimedia Broadcast Multicast Service MBMS-GW MBMS-gatewayMBSFN Multimedia Broadcast over an SFN MCE Multicell Coordination Entity MME Mobility Management EntityMPEG Moving Picture Experts Group

MPEG- MPEG-Dynamic Adaptive DASH Streaming over HTTPNBC National Broadcasting CompanyOFDM orthogonal frequency division multiplexing PGW packet data network gateway SDK software development kitSFN single-frequency networkSGW service gatewaySNR signal-to-noise ratioTDD time division duplexUDP User Datagram ProtocolUE user equipment

The solution is built on LTE technology, extending the LTE/EPC with an efficient point-to-multipoint distribution feature that can serve many eMBMS-capable LTE devices with the same content at the same time. It can be used to boost capacity for live and on-demand content so that well-liked websites, breaking news or popular on-demand video clips can be broadcast – off-loading the network and providing users with a superior experience.

Single-frequency network (SFN) tech-nology is used to distribute broadcast streams into well-defined areas – where all contributing cells send the same data during exactly the same radio time slots. The size of the coverage area of an LTE SFN can vary greatly, from just a few cells serving a stadium, to many cells delivering content to an entire coun-try. eMBMS-enabled devices can select

4


Capture your audience

solution. The concept has been built on eMBMS technology and based on a set of use cases that can be divided into two main categories

delivery of live premium content; andunicast off-loading (for example, local device caching).

Premium contentDespite the diversity of available con-tent and an obvious shift by subscribers towards on-demand viewing, watching certain events and programs live contin-ues to appeal to large audiences.

London 2012 is a good example of an event that enjoyed widespread live-viewing appeal. Ratings place the NBC coverage of the games as some of the most watched TV in US history; almost half of the online video streams were delivered to tablets or smartphones, and revenue expectations were far sur-passed. Some use cases for premium content follow.

Regional and localThis use case covers regional and local interest events, such as concerts, sports fixtures or breaking news. Such as the Super Bowl, FIFA World Cup matches, as well as elections and royal weddings. Given suitable content security and dig-ital-rights handling, this use case can be enhanced to allow users to store and replay the event on-demand from their device for a certain period of time.

Venue castingThis use case covers specific locations such as shopping malls, museums, air-ports, university campuses and amuse-ment parks. In this case, the operating enterprise may wish to broadcast con-tent to users, which can vary from breaking news of national interest to very specific information such as special offers available at the mall, additional information about the main artist of an art exhibition, or departures and arrivals information at the airport.

For all of these premium-content use cases, operators can deliver services on a nationwide basis as well as locally. The duration of a broadcast and the size of the geographical area where it is available can be managed dynami-cally, depending on the nature and rel-evance of the content. By using unicast for blended services, broadcast services can be complemented with interactiv-ity – opening up new ways to generate revenue from content.

At a soccer match, for example, these value-added services could include video streams carrying footage from additional camera angles, diverse audio coverage and live results of related matches taking place at the same time in other stadiums.

Unicast off-loadingMBMSs are traditionally associated with the delivery of live, linear TV, although

this technology also supports file deliv-ery. Exploiting this and the caching capability available in both mobile and fixed devices creates new possibilities for a range of use cases.

Popular contentOperators can choose to deliver popu-lar TV and video clips to the local cache of a user’s device at their convenience. Based on content popularity and busy-hour-traffic distribution, operators can deliver content when network load is low. Content shared on popular vid-eo streaming sites, as well as the con-tent provided by national and cable TV channels can all be pre-loaded to mobile devices through broadcast – sig-nificantly reducing the overall network capacity required to deliver frequently-consumed video streams.

News Daily clips and subscription content such as a magazine can be pre-delivered to the cache of a subscriber’s preferred device for that content.

Software upgradesUpgrades to application software and operating systems are usually released over the network to large numbers of subscribers at the same time. This tra-ditional way of performing an upgrade can be a burden on the network. By using LTE Broadcast instead, upgrades can be distributed as packages to a multitude of devices at little expense in terms of required resources – an approach that is particularly advanta-geous if the broadcast can be delivered during off-peak hours.

X %

Broadcast

FIGURE 1 Broadcast versus unicast

Y%,Y%

Y%, Y%

Y%, Y%

Unicast

Table 1: Broadcast versus unicast

Broadcast One data channel per content

Limited data channels, unlimited number of users

Resource allocation viewer independent

Unicast One data channel per user

Unlimited channels, limited number of users

Resources allocated when needed

5


the radio-channel quality and overall traffic volumes within the cell.

Broadcast is implemented as an extension to the existing EPS architec-ture (see Figure 7 and Box B). Ericsson’s LTE Broadcast system is mainly a soft-ware upgrade applied to existing nodes. The concept was designed according to 3GPP MBMS 23.246 for E-UTRAN and to coexist with unicast-data and voice services.

LTE Broadcast gives operators the flexibility to tailor the way content is delivered to suit their capabilities.

Service dynamicsThis supports live streaming and file-delivery use cases. Different service combinations may be delivered simul-taneously over the same bearer.

Time dynamics LTE Broadcast activation triggers the allocation of radio resources on a needs basis. A session may be active for a short time say several minutes or for longer periods: several days in some cases. When the session is no longer active, the assigned radio and system resources can be reallocated for use by other services.

Location dynamicsLTE Broadcast can be activated for small geographical locations, such as stadi-ums and city centers, or for large areas, covering say an entire city or region. As long as there is sufficient capacity in the network, multiple broadcast ses-sions can be active simultaneously.

Resource allocation dynamicsThis involves the free allocation of resources for LTE Broadcast. Up to 60 percent of the FDD radio resources and up to 50 percent for TDD can be assigned to a broadcast transmission.

Principles of the radio interfaceThe LTE radio interface is based on OFDM in the downlink, where the fre-quency selective wideband channel is subdivided into narrowband chan-nels orthogonal to each other. In time domain, a 10ms radio frame consists of subframes of 1ms each; where a sub-frame is the smallest unit with full fre-quency domain that can be allocated to a broadcast transmission.

With eMBMS, all users within the

M2M and B2BOver the coming decade, machine-to-machine (M2M) data traffic and the internet of things will create more connectivity demands on the network and create the need for diverse types of eMBMS LTE-enabled devices. LTE Broadcast technology supports efficient one-to-many transfer of machine data in any file format, which can be used for M2M use cases, off-loading the network and providing the essential machine connectivity and control.

Ericsson value propositionThe concept of Ericsson’s LTE Broadcast solution enables unicast and broadcast service blending, aiming to help meet the challenges created by rising mobile usage and the growth of video traffic in LTE networks. The solution covers the entire chain from live encoder, through delivery via point-to-multipoint trans-port to devices.

Particular focus has been placed on the specification and implementation of the device, starting with the physical chipset as well as transport control mid-dleware – essential enablers for the cre-ation and deployment of eMBMSs.

Implementing live streaming with MPEG-DASH4 is a technology choice that

supports the common use of a player on devices and a live encoder head-end system for both unicast and broadcast – reducing operating costs and maxi-mizing infrastructure usage. As out-lined later in this article, extensive simulation, lab testing and field trials have been conducted with the aim of characterizing the spectral efficiency of eMBMSs in deployed networks with mixed traffic profiles.

The results show that live video broad-cast with commercially acceptable lev-els of video and audio degradation is achievable. For video broadcasting to smartphones and tablets, compression using the H.2645 standard is feasible, with HEVC6 coming sometime in the near future.

System architectureBroadcast and unicast radio channels coexist in the same cell and share the available capacity. The subset of avail-able radio resources can temporarily be assigned to a broadcast radio channel.

Mobile-communication systems such as LTE are traditionally designed for uni-cast communication, with a separate radio channel serving each device. The resources allocated to the device depend on the data rate required by the service,

Sector edgemultipath gain

Maximum usable set of subframes

Radio frame = 10ms

Subframe = 1ms

bandwidth

0 1 2 3 4 5 6 7 8 9 0 1

t

Cell C1

C1C2

Cell C2

FIGURE 2 SFN principles

6



broadcast area, provided they have the right subscription level and an MBMS-capable device, can receive broadcast-ed content. By setting up a single bearer over the radio interface, operators can distribute a data stream to an unlimit-ed number of users.

Although it is possible to deliver broadcasts within a single cell, the con-cept becomes truly interesting with SFN, the principles of which are illus-trated in the lower part of Figure 2.

Broadcast data is sent over synchro-nized SFN – tightly synchronized, identical transmissions from multiple cells, using the same set of subframes and modulation and coding schemes, appear to the device as a transmission from a single large cell over a time-dis-persive channel. This improves received signal quality and spectral efficiency (as shown in Figure 2). For a more detailed description, refer to LTE/LTE-Advanced for Mobile Broadband7.

The maximal usable set of subframes is shown in the top left of the diagram, and the nodes are time-synchronized to a high precision.

By using long data-symbol duration in OFDM, it is possible to mitigate the effect of inter-symbol interference (ISI) caused by delayed signals. For addi-tional protection against propagation delays LTE/OFDM uses a guard interval – delayed signals arriving during the guard interval do not cause ISI and so the data rate can be maintained. For SFN, unlike unicast, signals arrive from many geographically separate sources and can incur large delay spread.

Consequently, one of the factors lim-iting MBMS capacity is self-interference from signals from transmitters with a delay that is greater than the guard interval (low transmitter density). To overcome this, a long cyclic prefix is added to MBSFN-reserved subframes to allow for the time difference in the receiver and corresponds to an ISD of approximately 5km.

Architecture The eMBMS architecture, shown in Figure  3, is designed to handle trans-mission requirements efficiently.

The Broadcast Multicast Service Center (BM-SC) is a new network ele-ment at the heart of the LTE Broadcast-distribution tree. Generic files or

MPEG-DASH live video streams are carried as content across the BM-SC and made available for broadcast. The BM-SC adds resilience to the broadcast by using AL-FEC – which adds redun-dancy to the stream so that receivers can recover packet losses – and supports the 3GPP-associated delivery proce-dures. These procedures include uni-cast base file repair – allowing receivers to fetch the remaining parts of a file through unicast from the BM-SC and reception reporting, so operators can collect QoE reports and make session-quality measurements.

Another new network element is the MBMS-GW, which provides the gateway function between the radio and service networks. It forwards streams from the BM-SC to all eNBs participating in the SFN transmission. IP multicast is used on the M1 interface between the gate-way and the eNBs, so that the packet rep-lication function of existing routers can be used efficiently. The gateway routes MBMS session control signaling to the MMEs serving the area. The MMEs in turn replicate, filter and forward ses-sion control messages to the eNBs partic-ipating in the specific broadcast session.

The eNBs provide functionality for configuration of SFN areas, as well

as broadcasting MBMS user data and MBMS-related control signaling on the radio interface to all devices. Note, the eNB contains the 3GPP Multicell Coordination Entity (MCE) function.

eMBMS LTE-enabled devices are an essential part of the ecosystem. LTE capabilities are becoming inte-grated into more and more types of devices and may be implemented on devices other than phones and tablets such as embedded platforms for M2M communications.

The UE platform is divided into three main blocks (see Figure  4):

the lower block incorporates the LTE radio layers, which are typically implemented in the LTE chipset, supporting unicast as well as broadcast; the middleware block handles the FLUTE protocol8, AL-FEC decoding, unicast file repair and other functions. It includes transport control functions, such as service scheduling, as well as a cache for post-broadcast file processing; andthe top platform block exposes APIs to the middleware and connectivity layer methods.

Application development is enabled through an SDK, which provides the platform APIs. The SDK enables developers to create and test

Unicast

S1-U

S11

Sm SGi

SGmb

SGi-mb

M3 / S1-MME

M1

S/PDN-GW

BM-SC

MBMS-GW

eNB

eNB

eNB

MME

ContentControlUser data

FIGURE 3 Architecture – with only eMBMS components shown

7


eMBMS-enabled applications with-out requiring detailed knowledge of the underlying transport, control, or radio-bearer technology.

Spectral efficiencyAccording to 3GPP specifications, eMBMSs and unicast services should be provisioned on a shared frequency. Consequently, while a broadcast service is active, radio-interface resources can be borrowed from unicast capacity.

Spectral efficiency can be defined as the possible information rate trans-mitted over a given bandwidth with a defined loss rate. The information loss rate depends on the modulation and coding scheme used for physical trans-missions and the protection offered by AL-FEC. This definition of spectral effi-ciency includes packet overheads, such as AL-FEC redundancy.

The simulation results from an evalu-ation of spectral efficiency are shown in

Figure  5. The results associated with a broadcast transmission depend on the ISD in a link budget – signal-to-noise ratio (SNR) – limited deployment. Two urban environments were simulated: indoor scenarios with 20dB penetration loss and in-car scenarios with 6dB loss assuming 95 percent coverage probabil-ity in all cases.

The failure criterion used was 10-3 BLER (corresponding to a packet loss of four packets per hour) and simulations were run with and without AL-FEC. An ideal Raptor code with FEC covering 2s per source block was used in this evalu-ation. The payload for each source block consisted of 50 packets with each IP packet spanning two transport blocks.

The MBSFN simulation area -included 19 sites, each with three sectors. The results show that MBMS spectral effi-ciency of about 1-3bps/Hz (indoor/in-car) could be achieved for a cellular ISD of up to 2km. The simulation results and addi-tional testing show that FEC improves video quality and saves capacity.

From the graphs in Figure 5, it is pos-sible to conclude that when ISD is less than 1km, spectral efficiency is great-er than 2.5b/s/Hz. By allocating one sub-frame for MBMS transmission in 20MHz spectrum, corresponding to 10 percent of capacity, the achievable data rate is in the range of 5Mbps.

Live video and file deliveryThe two main eMBMS use cases are live streaming and on-request file delivery. Live streaming supports services for real-time video and audio broadcast-ing, and on-request file delivery enables services such as unicast off-load (local device caching), software updates and M2M file loading. In fact, any arbitrary file or sequence of files can be distribut-ed over eMBMSs.

The target broadcast area for these use cases may be any desired size – some scenarios require a small broadcast area, such as a venue or a shopping mall, and other cases require much larger areas, even up to nationwide coverage. Ericsson has selected MPEG-DASH for live streaming delivery over eMBMSs. This solution slices the live stream into a sequence of media segments, which are then delivered through the system as independent files.

Typically, HTTP is used to fetch these

——

0 1 2 3 4 5 6 7 8 90

0.5

1.0

1.5

2.0

2.5

3.0

ISD (km)

Spectral efficiency (b/s/Hz)

w/o AL-FECAL-rBLER=1e-3AL-rBLER=1e-5

indoor; - - - in-car

FIGURE 5 Evaluating spectral efficiency

SGmb

SGi-mb

BM-SC

Platform APIs

eMBMSmiddleware

LTE chipset(L1, L2, L3)

User equipment

SDK

Appdownload

LTEBroadcast

and unicastnetwork

Content

FIGURE 4 UE and SDK in eMBMS ecosystem

 BOX B   Standards The standard-ization of MBMS started in 3GPP with Rel-6, which supported GERAN and UTRAN access networks. Over time, 3GPP has improved the access network support by, for example, defining the integrated mobile broad-cast (IMB) solution, which uses UTRAN TDD bands to offer up to 512kbps per content channel. Support for E-UTRAN access (LTE) was added to 3GPP Rel-9 as part of the eMBMS standardization activity.

8



files. In the eMBMS case, one quality rep-resentation is delivered as a sequence of files through eMBMSs using MBMS file delivery.

By using MPEG-DASH with eMBMSs, the same live encoder and common clients can be used for unicast and broadcast offerings. This solution also supports using the same system proto-col stack for both live streaming and file-delivery implementation.

The IETF FLUTE protocol8 allows dis-tribution of files over unidirectional links using UDP. Most service-layer fea-tures can be used for both streaming and file delivery; transmission reliabili-ty can be increased using AL-FEC in both cases. File delivery can also make use of the unicast file-repair feature – allowing UEs to fetch any missing file segments. However, this feature is not intended for use with services that have real-time requirements, such as live streaming.

With FLUTE, delivery and eMBMS sessions are used, where the duration of a delivery session may span one or more eMBMS sessions. The broadcast is active for the entire eMBMS session, during which UEs can receive content. The relationship between delivery ses-sions and eMBMS sessions is shown in Figure  6. Service announcement is used to inform devices about delivery sessions and also about eMBMS sessions using a schedule description. UEs do not need to monitor the radio interface for eMBMS sessions continuously.

In Figure 6, the schedule description instructs the UE to expect an eMBMS session between t2 and t3 and between t6 and t7. Before the UE expects an eMBMS session, it is already active on the radio interface (t1 < t2). When it comes to file-delivery services, it is preferred that devices should search for sessions prior to expected transmission time on the radio, to ensure that they do not miss the start of a transmission.

The example in Figure 6 could repre-sent a service, such as downloading an application that allows users to activate, receive and interact with the broadcast using unicast services from a phone, tab-let or television.

From the point of view of the user and the UE middleware, the two broad-casts belong to the same MBMS user ser-vice, which presents a complete offering including activation and deactivation.

ConclusionsThe data volume in mobile networks is booming mostly due to the success of smartphones and tablets.

LTE Broadcast is one way of provid-ing new and existing services in areas that can at times be device dense, such as stadiums and crowded city centers. Single-frequency network technology

Service announcementinforms the UE aboutthe schedule

t2 to t3UEs expects to receive dataof that FLUTE session

t6 to t7UEs expects to receive dataof that FLUTE session

Service announcementinforms the UE aboutthe schedule

eMBMS session

Service announcement

Delivery session (FLUTE)

eMBMS session

t1 t2 t3 t4 t5 t6 t7 t8 Time

FIGURE 6 Example of two scheduled broadcasts

is used to distribute the content over the air interface.

LTE Broadcast provides operators with techniques to deliver consistent service quality, even in highly crowded areas. Such techniques for delivering content efficiently are valuable as they free up capacity, which can be used for other services and voice traffic.

App

FR/RR (HTTP) HTTP

SGiS5/S8

S1-US11

S1-MME

M3

M1

SmUu

SGmb

SGi-mb

UE eNB MBMS-GW

SGW PGW

BM-SC

Applicationserver

CDN/live encoder

MME

FIGURE 7 eMBMS architecture

9


Michael Slssingar

is an Ericsson senior specialist in service delivery architectures and holds a post-graduate

diploma and master’s in computing and software engineering. He has held many senior engineering roles at Ericsson, mainly in the media-delivery field, and has contributed to the Ericsson IPTV and Mobile TV delivery solutions. In the field of MBMS, Slssingar initially specialized in WCDMA MBMS, where he helped develop the Ericsson Content Delivery System. More recently, he has worked with LTE eMBMS broadcast, where he has a strong interest in the service layer BM-SC node, UE middleware and metadata provisioning areas.

Thorsten Lohmar

joined Ericsson in Germany in 1998 and worked in different Ericsson Research units

for several years. He worked on a variety of topics related to mobile-communication systems and led research projects specifically in the area of multimedia technologies. On the development front, he is focusing on the technical coordination of eMBMSs with an end-to-end perspective. He is currently working as a senior specialist for end-to-end video delivery, principally in mobile networks. Lohmar holds a Ph.D. in electrical engineering from RWTH Aachen University, Germany.

1. Ericsson,2012,EricssonConsumerLabreport,TVand video – changing the game, available at: http://www.ericsson.com/res/docs/2012/consumerlab/consumerlab-tv-video-changing-the-game.pdf

2. Ericsson, November 2012, Ericsson Mobility Report, On the pulse of the Networked Society, available at: http://www.ericsson.com/res/docs/2012/ericsson-mobility-report-november-2012.pdf

3. MobileContentVenture,June2012,DyleMobileTVDataReport, available at: http://www.dyle.tv/assets/Uploads/DyleReport.pdf

4. ISO/IEC 23009-1:2012, Information technology – Dynamic adaptive streaming over HTTP (DASH) — Part 1: Media presentation description and segment formats, available at: http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_detail.htm?csnumber=57623

5. ITU-T H.264 Advanced video coding for generic audiovisual services, available at: http://www.itu.int/rec/T-REC-H.264

6. ITU-TH.265/ISO/IEC23008-2HEVC,availableat: http://www.itu.int/ITU-T/aap/AAPRecDetails.aspx?AAPSeqNo=2741

7. ErikDahlman,StefanParkvall,JohanSköld,2011, LTE/LTE-Advanced for Mobile Broadband, available at: http://www.elsevier.com/books/4g-lte-lte-advanced-for-mobile-broadband/dahlman/978-0-12-385489-6

8. IETF RFC 3926, FLUTE – File delivery over unidirectional transport, T. Paila, et al., October 2004, available at: http://tools.ietf.org/html/rfc3926

References

Vera Kenehan

is a strategic product manager within LTE Radio and has worked with several generations of

radio-access technologies, including LTE, WCDMA and PDC. She was largely involved in the initial standardization of LTE, including eMBMS. For the past two years, she has been working on the MBMS product as well as promoting and bringing eMBMS to the market. She holds a master’s in telecom engineering from the University of Belgrade, Serbia.

Stig Puustinen

is a senior project manager at System Management within Business Unit Networks,

where he is currently running an LTE/EPC systems project involving extensive eMBMSs work. He joined Ericsson in 1991, and has since held a variety of project- and program-management roles. He was involved in the early releases of GSM, the first introduction of WCDMA/HSPA and the first release of LTE/EPC.

10



Nine decades of innovationAutomatic exchanges to smart networks

The history of our technology is deeply entrenched in that of the telecoms industry. War, international terrorism, and developments in other industries such as railways and the more recent digital revolution have shaped our playing field.

Our technical expertise is not just based on theory; it is about how to apply the right technology to create commer-cially viable products. Standards – and the ability to create and evolve them – have helped us and our industry become global providers of interoperable solutions – a cornerstone of the Networked Society.

Over the past nine decades, the world has probably changed more than ever before. Here are a few highlights from Ericsson’s history and some of the world events that have played their part in the evolution of telecoms.

In 1924, Ericsson launched its automatic 500-point sys-tem. Highlighted in the very first issue of Ericsson Review, this revolutionary switching system was demonstrated at the Gothenburg Exhibition of 1923. In hindsight, it sounds curious that attendees could ‘watch’ the switching process, which at the time was mechanical. By the 1930s, Ericsson had delivered about 100 systems with a total of more than 350,000 lines. Sales of the system continued to rise over the coming decades, not declining significantly until the 1970s. By 1974, 4.8 million lines using this system were in operation in pub-lic telephone stations.

The very first 500-point system was put into service at the Norra Vasa exchange in Stockholm, and was still in operation 60 years later. At the time of installation in 1924, Televerket – the Swedish government agency for telecommu-nications (1853-1993) – had four different systems to choose from, including a crossbar system also developed by Ericsson. Televerket’s choice was a decisive one for the future develop-ment of Ericsson.

In the 1920s and 1930s, the opening of a new telephone exchange was a major event for small towns and villages. Official opening ceremonies were often carried out by the local mayor, to the backdrop of a brass band and refreshments provided for the locals by the operator and the vendor.

In the 1930s, Ericsson introduced a photoelectric announc-ing machine – a simple device that could deliver short prere-corded messages. This offloaded the work of operators and had the ability to deliver longer announcements, such as the speaking clock and – later in the decade – automatic weather forecasts. The first device for weather forecasts was put into service in Stockholm on June 1, 1936 and was the first of its kind in the world. It wasn’t until two decades later that they were replaced with newer technologies.

In the 1940s, Ericsson introduced an information manage-ment structure called the ABC System. Ericsson still uses this system to classify and manage its information and products.

Continued on page 19...


11

Re:view

The Roaring TwentiesIn 1924, in the very first

issue of Ericsson Review, the editor stated that the objective of the magazine was to take up points of design and construction, which had not yet reached final standardization, for discussion. The cover article featured Ericsson’s presence at the Gothenburg Exhibition (1923), where a giant replica of our standard table set telephone housed a complete Ericsson exchange for 500 lines, to which a few telephones were connected. Visitors could make calls and watch the switching process, through the plate-glass windows.

Depression and modernismThe cover of issue 2, 1934

shows the ‘photo-electric talking machine for automatic time indication.’ This issue included an account of a carrier system supplied by Ericsson to the Indian Radio and Communications Company. In the spirit of modernism, this issue included a number of articles relating to the use of clocks and timing mechanisms in industry. An article on party lines discussed the problem of connecting several telephone instruments to one line. This was of particular interest at the time, as exchanges in rural districts were being automated and party lines were in need of some degree of modification. One of the main concerns was how to use party lines without altering the subscriber equipment.

The war years Ericsson Review was not published in English between

1940 and 1944 due to the ongoing world war. Apart from the fact that the paper situation necessitated a reduction both of quantity and quality of printing paper, the journal was issued as usual in its Swedish edition during the war years. In 1945, a collection of some of the articles published during the war in Swedish were printed in a composite English language edition.

Ericsson Review, issues 1 & 2, 1924.

Ericsson Review, issue 2, 1934.

Non-line-of-sight microwave backhaul for small cellsThe evolution to denser radio-access networks with small cells in cluttered urban environments has introduced new challenges for microwave backhaul. A direct line of sight does not always exist between nodes, and this creates a need for near- and non-line-of-sight microwave backhaul.

solution that is also effective for small-cell backhaul (see Figure 1).

Network architects aim to dimen-sion backhaul networks to support peak cell-capacity3 – which today can reach 100Mbps and above. However, in reality, there is a trade-off among cost, capacity and coverage resulting in a backhaul solution that, at a minimum, can support expected busy-hour traffic with enough margin to account for sta-tistical variation and future growth: in practice around 50Mbps with availabil-ity requirements typically relaxed to 99-99.9 percent. Such availability levels require fade margins of the order of just a few decibels for short-link distances.

For small-cell backhaul simplicity and licensing cost are important issues. Light licensing or technology-neutral block licensing are attractive alterna-tives to other approaches such as link licensing, as they provide flexibility4. Using unlicensed frequency bands can be a tempting option, but may result in unpredictable interference and degrad-ed network performance. The risk associated with unlicensed use of the 57-64GHz band is lower than that asso-ciated with the 5.8GHz band, owing to higher atmospheric attenuation, sparse initial deployment, and the possibility of using compact antennas with nar-row beams, which effectively reduces interference.

Providing coverage in locations with-out a clear line of sight is a familiar part of the daily life of mobile-broadband and Wi-Fi networks. However, maybe because such locations are common-place, a number of widespread myths and misunderstandings surrounding NLOS microwave backhaul exist – for example, that NLOS microwave back-haul needs sub-6GHz frequencies, wide-beam antennas and OFDM-based radio

Complementing the macro-cell layer by adding small cells to the RAN introduc-es new challenges for backhaul. Small-cell outdoor sites tend to be mounted 3-6m above ground level on street fix-tures and building facades, with an inter-site distance of 50-300m. As a large number of small cells are necessary to support a superior and uniform user experience across the RAN2, small-cell backhaul solutions need to be more cost-effective, scalable, and easy to install than traditional macro backhaul tech-nologies. Well-known backhaul tech-nologies such as spectral-efficient LOS microwave, fiber and copper are being tailored to meet this need. However, owing to their position below roof height, a substantial number of small cells in urban settings do not have access to a wired backhaul, or clear line of sight to either a macro cell or a remote fiber backhaul point of presence.

The challenges posed by locations without a clear line of sight are not new to microwave-backhaul engineers, who use several established methods to overcome them. In mountainous terrain, for example, passive reflectors and repeaters are sometimes deployed. However, this approach is less desir-able for cost-sensitive small-cell back-haul, as it increases the number of sites. In urban areas, daisy chaining is often used to reach sites in tricky locations – a

JONA S H A NSRY D, JONA S E DSTA M, BE NGT-E R I K OL SSON A N D CH R I ST I NA L A R SSON

BOX A Terms and abbreviationsFDD frequency division duplexingLOS line-of-sightMIMO multiple-input, multiple-output

NLOS non-line-of-sightOFDM orthogonal frequency division multiplexing RAN radio-access networkTDD time division duplexing

Using non-line-of-sight (NLOS) propagation is a proven approach when it comes to building RANs However, deploying high-performance microwave backhaul in places where there is no direct line of sight brings new challenges for network architects. The traditional belief in the telecom industry is that sub-6GHz bands are required to ensure performance for such environments. This article puts that belief to the test, providing general principles, key system parameters and simple engineering guidelines for deploying microwave backhaul using frequency bands above 20GHz. Trials demonstrate that such high-frequency systems can outperform those using sub-6GHz bands – even in locations with no direct line of sight. Point-to-point microwave is a cost- efficient technology for flexible and rapid backhaul deployment in most locations. It is the dominant backhaul medium for mobile networks, and is expected to maintain this position as mobile broadband evolves; with micro-wave technology that is capable of pro-viding backhaul capacity of the order of several gigabits-per-second1.

12


Dispelling the NLOS myths

technologies to meet coverage and capacity requirements. Despite this, a number of studies on NLOS transmis-sion using frequency bands above 6GHz, for example, have been carried out for fixed wireless access5 and for mobile access6. Coldrey et al. showed that it is realistic to reach 90 percent of the sites in a small-cell backhaul deploy-ment with a throughput greater than 100Mbps using a paired 50MHz chan-nel at 24GHz7.

NLOS principlesAs illustrated in Figure 1, all NLOS prop-agation scenarios make use of one or more of the following effects:

diffraction; reflection; andpenetration.

All waves change when they encounter an obstacle. When an electromagnetic wave hits the edge of a building, dif-fraction occurs – a phenomenon often described as the bending of the signal. In reality, the energy of the wave is scat-tered in the plane perpendicular to the edge of the building. The energy loss – which can be considerable – is propor-tional to both the sharpness of the bend and the frequency of the wave8.

Reflection, and in particular random multipath reflection, is a phenomenon that is essential for mobile broadband using wide-beam antennas. Single-path reflection using narrow-beam antennas is, however, more difficult to engineer owing to the need to find an object that can provide the necessary angle of inci-dence to propagate as desired.

Penetration occurs when radio waves pass through an object that completely or partially blocks the line of sight. It is a common belief that path loss result-ing from penetration is highly depen-dent on frequency, which in turn rules out the use of this effect at higher fre-quencies. However, studies have shown that in reality path loss due to penetra-tion is only slightly dependent on fre-quency, and that in fact it is the type and thickness of the object itself that creates the impact on throughput9, 10. For exam-ple, thin, non-metallic objects – such as sparse foliage (as shown in Figure 1) –add a relatively small path loss, even for high frequencies.

Deployment guidelines can be defined given a correct understanding

and application of these three propaga-tion effects, giving network engineers simple rules to estimate performance for any scenario.

System propertiesA simplified NLOS link budget can be obtained by adding an NLOS path atten-uation term (ΔLNLOS) to the traditional LOS link budget, as shown in Equation 1.

Equation 1 PRX = PTX + GTX + GRX - 92 - 20log(d) - 20log(f) - LF - ΔLNLOS

Here, PRX and PTX are the received and transmitted power (dBm – ratio of pow-er in decibels to 1 milliwatt); GTX and GRX are antenna gain (in decibels isotro-pic – dBi) for the transmitter and receiv-er respectively; d is the link distance (in kilometers); f is the frequency (in gigahertz); LF is any fading loss (in deci-bels); and ΔLNLOS is the additional loss (in decibels) resulting from the deploy-ment of NLOS-propagation effects. Not shown in this equation is the theoreti-cal frequency dependency of the anten-na gain, which for a fixed antenna size will increase as 20log(f) and as a conse-quence, the received signal – PRX – will

actually increase as 20log(f) when car-rier frequency is increased for a fixed antenna size. This relationship indi-cates the advantage of using higher fre-quencies for applications where a small antenna form factor is of importance – as is the case for small-cell backhaul.

To determine the importance of NLOS-system properties, Ericsson car-ried out measurement tests on two com-mercially available microwave backhaul systems in different frequency bands (described in Table 1). The first system used the unlicensed 5.8GHz band with a typical link configuration for applica-tions in this band. The air interface used up to 64QAM modulation in a 40MHz-wide TDD channel with a 2x2 MIMO (cross-polarized) configuration pro-viding full duplex peak throughput of 100Mbps (200Mbps aggregate). The sec-ond system, a MINI-LINK PT2010, used a typical configuration for the licensed 28GHz band, based on FDD, 56MHz channel spacing and single-carrier technology with up to 512QAM modu-lation, providing full duplex through-put of 400Mbps (800Mbps aggregate). To adjust the throughput based on the quality of the received signal, both

FIGURE 1 Microwave backhaul scenarios for small-cell deployment

Daisy chain Penetration Reflection Diffraction

Fiber

Table 1: Test system specifications

SYSTEM TECHNOLOGY CHANNEL SPACING

ANTENNA GAIN

OUTPUTPOWER

PEAKTHROUGHPUT

5.8GHz TDD/OFDM 64QAM

40MHz 17dBi 19dBm 100Mbps

28GHz FDD/single carrier 512QAM

56MHz 38dBi 19dBm 400Mbps

13


Here, the margin is defined as the difference between received power (according to Equation 1) and the receiv-er threshold for a particular modula-tion level (throughput) – in line of sight conditions without fading (Lf = 0). If ΔLNLOS caused by any NLOS effect can be predicted, the curves in Figure 2 can be used to estimate throughput. The

advantages of using higher frequencies are clear: with comparable antenna siz-es, the link margin is about 20dB higher at a peak rate of 400Mbps for the 28GHz system compared with the 5.8GHz sys-tem at a peak rate of 100Mbps.

Measurements DiffractionIt is commonly believed that the dif-fraction losses occurring at frequencies above 6GHz are prohibitively high, and consequently, deploying a system using this effect for NLOS propagation at such frequencies is not feasible. However, even if the absolute loss can be relative-ly high, 40dB and 34dB for the 28GHz and 5.8GHz systems respectively (with a diffraction angle of 30 degrees), the relative difference is only 6dB8 – much less than the difference in gain for com-parable antenna sizes even when tak-ing into account the higher free-space loss for the 28GHz system (see Figure 2).

Figures 3A and 3B show the set-up and measured results of a scenario designed to test diffraction. A first radio was positioned on the roof of an office building (marked in Figure 3A with a white circle). A second radio was mount-ed on a mobile lift, placed 11m behind a 13m-high parking garage. The effect on the signal power received by the sec-ond radio was measured by lowering the mobile lift. Figure 3B shows the mea-sured received signal-power versus dis-tance below the line of sight for both test systems, as well as the theoretical received power calculated using the ide-al knife-edge model8. Both radios trans-mitted 19dBm output power, but due to the 21dBi lower antenna gain for the 5.8GHz system, the received signal for this radio was 20dB weaker after NLOS propagation than the 28GHz system.

The measured results compare well against the results based on the theoreti-cal model, although an offset of a couple of decibels is experienced by the 28GHz system – a small deviation that is expect-ed due to the simplicity of the model.

To summarize, diffraction losses can be estimated using the knife-edge model8. However, due to the model’s simplicity, losses calculated by it are slightly underestimated. This can be compensated for in the planning pro-cess by simply adding a few extra deci-bels to the loss margin.

systems used adaptive modulation. Physical antenna sizes were similar, but due to the frequency dependency of the antenna gain and the parabolic type used in the 28GHz system, it offered a gain of 38dBi while the flat antenna of the 5.8GHz system reached 17dBi.

Link margin versus throughput and hop distance is shown in Figure 2.

Link distance (m)

Link margin (dB)

80

70

60

50

40

30

20

1050 100 150 200 250 300 350 400 450 500

28GHz5.8GHz

90Mbps

185Mbps

280Mbps10Mbps

60Mbps

80Mbps

100Mbps

400Mbps

FIGURE 2 Link margin as a function of throughput and distance

14



FIGURE 3A Test site for NLOS backhaul – diffraction

150m

© 2013 BLOM © 2013 Microsoft Corporation

The 28GHz system can sustain full throughput at much deeper NLOS than the 5.8GHz system, which is to be expected as it has a higher link mar-gin. Full throughput – 400Mbps – was achieved at 28GHz up to 6m below the line of sight, equivalent to a 30-degree diffraction angle, while the 5.8GHz sys-tem dropped to under 50Mbps at 3m below the line of sight. The link mar-gin is the single most important system parameter for NLOS propagation and, as expected, the 28GHz system performs in reality better in a diffraction scenar-io than a 5.8GHz system with compara-ble antenna size.

ReflectionThe performance characteristics of the 5.8GHz and 28GHz systems were mea-sured in a single-reflection scenario in an area dominated by metal and brick facades – shown in Figure 4A. The first radio was located on the roof of the office building (marked with a white circle), 18m above ground level; and the second on the wall of the same build-ing, 5m above ground, facing the street canyon. The brick facade of the building on the other side of the street from the second radio was used as the reflecting object, resulting in a total path length of about 100m. The reflection loss will vary with the angle of incidence, which in this case was approximately 15 degrees, resulting in a ΔLNLOS of 24dB for the 28GHz system and 16dB for the 5.8GHz system – figures that are in line with ear-lier studies11. Reflection loss is strongly dependent on the material of the reflect-ing object, and for comparison purpos-es ΔLNLOS for a neighboring metal facade was measured to be about 5dB for both systems with similar angle of incidence.

To summarize, it is possible to cover areas that are difficult to reach using multiple reflections in principle. However, taking advantage of more than two reflections is in practice prob-lematic – due to limited link margins and the difficulty of finding suitably aligned reflection surfaces. ΔLNLOS pre-dictions for a single-facade reflection in the measured area can be expect-ed to vary between 5dB and 25dB at 28GHz and between 5dB and 20dB at 5.8GHz. The throughput for both sys-tems measured over 16 hours is shown in Figure 4B.

The 28GHz system shows a stable throughput of 400Mbps, while the throughput for the 5.8GHz system, with a much wider antenna beam, dropped from 100Mbps to below 70Mbps. These variations are to be expected owing to the fact that the wider beam expe-riences a stronger multipath. OFDM is an effective mitigation technolo-gy that combats fading, which will,

at severe multipath fading, result in a graceful degradation of throughput – as illustrated. However, the narrow antenna lobe at 28GHz, in combina-tion with the advanced equalizer of the high- performance MINI-LINK radio, effectively suppresses any multipath degradation, enabling the use of a sin-gle-carrier QAM technology for NLOS conditions – even up to 512QAM and 56MHz channel bandwidths.

FIGURE 3B Throughput and received power – diffraction

FIGURE 4A Test site for NLOS backhaul – reflection

100m

Distance below line of sight (m)

Throughput (Mbps) Received power (dBm)

700

600

500

400

300

200

100

0

-10

-20

-30

-40

-50

-60

-70

-80–2 0 2 4 6 8 10

Received power 28GHz

Received power 5.8GHz

Theoretical power 28GHz

Theoretical power 5.8GHz

Throughput 28GHz

Throughput 5.8GHz

15


© 2012 TerraItaly © 2013 Microsoft CorporationPictometry Bird’sEye © 2012 Pictometry International Corp.

shown in Figure 5A. The circle and tri-angle symbols indicate where the radio beams exit the foliage.

Measurements were carried out under rainy and windy weather con-ditions, resulting in variations of the NLOS path attenuation, as shown in the received signal spectra for the 28GHz radio link in Figure 5B. Under LOS con-ditions the amplitude spectrum enve-lope reached -50dBm. Consequently, the excess path loss for the single-tree (sparse foliage) scenario varied between 0 and 6dB when measured for 5 min-utes. In the double-tree (dense foliage) case excess path loss varied from 8dB

PenetrationAs with the case for NLOS reflection, the path loss resulting from penetra-tion is highly dependent on the mate-rial of the object blocking the line of sight. The performance of both test systems was measured in a scenario shown in Figures 5A and 5B. The send-ing and receiving radios were located 150m apart, with one tall sparse tree and a shorter, denser tree blocking the line of sight. The radio placed on the mobile lift was positioned to measure the radio beam first after penetration of the sparse foliage and then lowered to measure the more dense foliage, as

up to more than 28dB. A complementa-ry experiment showing similar excess path loss was carried out at 5.8GHz.

To summarize, contrary to popular belief, a 28GHz system can be used with excellent performance results using the effect of NLOS penetration through sparse greenery.

Deployment guidelinesSo far, this article has covered the key system properties for NLOS propaga-tion – diffraction, reflection and pene-tration – dispelling the myth that these effects can be used only with sub-6GHz frequencies. The next step is to apply the theory and the test results to an actual deployment scenario for microwave backhaul.

Table 2 shows the indicative through-put for each NLOS scenario, using the measured loss from the examples above together with the graphs in Figure 2.

A trial site, shown in Figure 6, was selected to measure the coverage for an NLOS backhaul deployment scenario. Four- to six-story office buildings with a mixture of brick, glass and metal facades dominate the trial area. The hub node was placed 13m above ground on the corner of a parking garage at the south end of the trial area. By using the measured loss in the diffraction, reflec-tion and penetration from the tests as a rule of thumb, an indicative through-put for each NLOS scenario has been taken from Figure 2 and summarized in Table 2.

The colored areas in Figure 6 show the line of sight conditions for the trial site: the green areas show where pure LOS exists; the yellow areas indicate the use of single-path reflection; the blue areas indicate diffraction; and the red areas show where double reflection is need-ed. Areas without color indicate either that no throughput is expected or that they are outside the region defined for measurement. Measurements were made within the region delineated by the dashed white lines. Referring to Table 2, it is expected that the 5.8GHz system will meet small-cell backhaul requirements (>50Mbps throughput) within a 250m radius of the hub; and the 28GHz system should provide more than 100Mbps full duplex throughput up to 500m from the hub. To test the actual performance, a receiver node

Time (hours)

Throughput (Mbps)

120

380

400

420

100

80

60

0 2 4 6 8 10 12 14 16

28GHz5.8GHz

FIGURE 4B Throughput over time – single reflection

= sparse foliage

= dense foliage

FIGURE 5A Test site for NLOS backhaul – penetration

16



was placed 3m above ground measuring full duplex throughput along the main street canyon and in the neighboring streets. On account of the wide antenna lobe of the 5.8GHz system, realignment was not needed for the hub antenna for measurement purposes. For the 28GHz system, realignment of the narrow antenna beam was needed at each mea-surement point – a fairly simple proce-dure even under NLOS conditions.

The actual values observed at each measurement point exceeded or matched the predicted performance lev-els in Table 2. Due to the lack of correct-ly aligned reflection surfaces, providing backhaul coverage using the double-reflection technique (the red areas of the trial area in Figure 6) was only pos-sible for a limited set of measurements. Multipath propagation, including the reflection effects created by vehi-cles moving along the street canyon, was significant for the 5.8GHz system, but resulted only in slightly reduced throughput in some of the more diffi-cult scenarios for the 28GHz system.

Summary In traditional LOS solutions, high sys-tem gain is used to support targeted link distance and mitigate fading caused by rain. For short-distance solutions, this gain may be used to compensate for NLOS propagation losses instead. Sub-6GHz frequency bands are proven for traditional NLOS usage, and as shown in this article, using these bands is a viable solution for small-cell backhaul. However, contrary to common belief, but in line with theory, MINI-LINK microwave backhaul in bands above 20GHz will outperform sub-6GHz sys-tems under most NLOS conditions.

The key system parameter enabling the use of high-frequency bands is the much higher antenna gain for the same antenna size. With just a few simple engineering guidelines, it is possible to plan NLOS backhaul deployments that provide high network performance. And so, in the vast amount of dedicat-ed spectrum available above 20GHz, microwave backhaul is not only capa-ble of providing fiber-like multi-gigabit capacity, but also supports high perfor-mance backhaul for small cells, even in locations where there is no direct line of sight.

Table 2: Indicative bitrate performance for different NLOS key scenarios

LOS SINGLE REFLECTION

DOUBLEREFLECTION

DIFFRACTION* PENETRATION***

5.8GHz

0-100m 100Mbps 100Mbps 10Mbps** 80Mbps 100Mbps

100-250m 100Mbps 80Mbps <10Mbps** 60Mbps 100Mbps

250-500m 100Mbps 60Mbps <10Mbps** 10Mbps** 80Mbps

28GHz




*30-degree diffraction angle; **not recommended for small-cell backhaul; ***sparse foliage or similar

FIGURE 5B Channel amplitude response – penetration

Frequency (GHz)

Received power (dBm)

–40

–50

–60

–70

–80

~6dB fluctuation

Maximum power over 5 minutesMinimum power over 5 minutes

29.32 29.34 29.36 29.38 29.40 29.42

Sparse foliage

Frequency (GHz)

Received power (dBm)

–40

–50

–60

–70

–80

>20dB fluctuation

Maximum power over 5 minutesMinimum power over 5 minutes

29.32 29.34 29.36 29.38 29.40 29.42

Dense foliage

Resolution bandwidth: 50kHz

FIGURE 6 NLOS backhaul trial area

17


500m

100m

250m

© 2013 BLOM © 2013 Microsoft Corporation

Legend

line-of-sight

doublereflection

singlereflection

diffraction

Jonas Hansryd

joined Ericsson Research in 2008 and is currently managing the microwave high-speed

and electronics group. He holds a Ph.D. in electrical engineering from the Chalmers University of Technology, Gothenburg, Sweden and was a visiting researcher at Cornell University, Ithaca, US, from 2003-2004.

Jonas Edstam

joined Ericsson in 1995 and is currently head of technology strategies at Product Line Microwave

and Mobile Backhaul. He is an expert in microwave radio-transmission networks focusing on the strategic evolution of packet-based mobile backhaul and RAN. He holds a Ph.D. in applied solid-state physics from Chalmers University of Technology, Gothenburg, Sweden.

1. Ericsson, 2011, Ericsson Review, Microwave capacity evolution, available at: http://www.ericsson.com/res/docs/review/Microwave-Capacity-Evolution.pdf

2. Ericsson, 2012, White Paper, It all comes back to backhaul, available at: http://www.ericsson.com/res/docs/whitepapers/WP-Heterogeneous-Networks-Backhaul.pdf

3. NGMNAlliance,June2012,WhitePaper,SmallCellBackhaul Requirements, available at: http://www.ngmn.org/uploads/media/NGMN_Whitepaper_Small_Cell_Backhaul_Requirements.pdf

4. Electronic Communications Committee (ECC), 2012, Report 173, Fixed service in Europe – current use and future trends post, available at: http://www.erodocdb.dk/Docs/doc98/official/pdf/ECCRep173.PDF

5. Seidel, S.Y.; Arnold, H.W.; 1995, Propagation measurements at 28 GHz to investigate the performance of local multipoint distribution service (LMDS), available at: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=502029

6. Rappaport,T.S.;YijunQiao;Tamir,J.I.;Murdock,J.N.;Ben-Dor, E.; 2012, Cellular broadband millimeter wave propagation and angle of arrival for adaptive beam steering systems (invited paper), available at: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6175397

7. Coldrey,M.;Koorapaty,H.;Berg,J.-E.;Ghebretensaé,Z.;Hansryd,J.;Derneryd,A.;Falahati,S.;2012,Small-cellwireless backhauling: a non-line-of-sight approach for point-to-point microwave links, available at: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6399286

8. ITU, 2012, Recommendation ITU-R P.526, Propagation by diffraction, available at: http://www.itu.int/rec/R-REC-P.526-12-201202-I

9. Anderson, C.R.; Rappaport, T.S.; 2004, In-building wideband partition loss measurements at 2.5 and 60 GHz, available at: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=1296643

10. Okamoto, H.; Kitao, K.; Ichitsubo, S.; 2009, Outdoor-to-Indoor Propagation Loss Prediction in 800-MHz to 8-GHz Band for an Urban Area, available at: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=4555266

11. Dillard, C.L.; Gallagher, T.M.; Bostian, C.W.; Sweeney, D.G.; 2003, 28GHz scattering by brick and limestone walls, available at: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=1220086

References

Christina Larsson

joined Ericsson Research in 2010. Her current focus area is microwave backhaul

solutions. She holds a Ph.D. in electrical engineering from Chalmers University of Technology, Gothenburg, Sweden, and was a post-doctoral researcher at the University of St. Andrews, St. Andrews, UK, from 2004-2006.

Bengt-Erik Olsson

joined Ericsson Research in 2007 to work on ultra-high-speed optical communication

systems. Recently he switched interest to wireless-technology research, and is currently working on NLOS backhaul applications for microwave links. He holds a Ph.D. in optoelectronics from Chalmers University of Technology, Gothenburg, Sweden.

The authors gratefully acknowledge the colleagues who have contributed to thisarticle:Jan-ErikBerg,MikaelColdrey,AndersDerneryd,UlrikaEngström,SorourFalahati,FredrikHarrysson,MikaelHöök,BjörnJohannisson, Lars Manholm and Git Sellin

Acknowledgements

18



Toward the end of this decade, Ericsson trialed a crossbar system, 30 years after these switches were first put into prac-tical operation.

The Second World War had a profound effect on people and business. Widespread misery, rations and a shortage of raw materials naturally led many companies to diversify their operations. But overall, the slowdown in growth and reduced level of information exchange among researchers led to a dip in development. World average telephone density in 1930 was 2 per 100 capita; by 1950 this figure had risen to 3, and in the two decades following that, subscriber density more than dou-bled – reaching a total of 7 in 1970.

In preparation for the 1952 Summer Olympics, Ericsson installed its first commercial crossbar system in Helsinki, Finland, in 1950. The decision to move to crossbar switching came about during the war as Ericsson was developing small-er exchanges for rural communities and enterprises. This technology, however, presented new challenges in terms of traffic engineering and dimensioning in particular. In his the-sis on a study of congestion in link systems, lifelong Ericsson employee Christian Jacobæus presented a way to calculate traffic capacity that was subsequently trialed and became a worldwide standard.

Computing was the next wave of technology to be adopt-ed by the telecoms industry. Parts of Ericsson’s AKE range of telephone exchanges were computer controlled, the key fea-ture being a Stored Program Controlled (SPC) element, which managed the switches and operated in real time. The initial commercial deployments in 1968 were the first computer-controlled exchanges outside the US.

By the end of the 1960s, however, it had become clear that something different was needed. Something that was more flexible – a modular system that could be expanded to accommodate new technologies and services without the need for fundamental system changes: something that was future-proof.

The downturn of the global economy following the oil cri-sis in the 1970s again led to several tough years for industry in general. For Ericsson, a boost came when the Saudi Ministry of Telecommunications chose Ericsson AXE exchanges in what was to be the biggest contract in the history of telecoms at that time. This digital exchange introduced modular design and became one of the world’s most successful switching systems.

In 1984, Ericsson Review published a special ‘F’ edition dedicated to fiber optic. This issue covered every aspect of fiber optics from cable manufacture to installation, applica-tions and device technology. Ericsson’s transmission expertise goes right back to the late 1920s when it began manufactur-ing loading coils, early signs of Ericsson’s ethos to provide the telecoms industry with a wider range of products and services.

Continued on page 47...



19

Re:view

Recovery and prosperity In 1954, the cover article

of Ericsson Review issue 1 looked at the traffic reliability of the crossbar system, which Ericsson delivered to the Helsinki Telephone Corporation in 1950. The cover illustrates the interior of the Helsinki exchange for PBX subscribers. The final testing of the system was carried out in the latter half of 1953 and gave a fault rate of 0.090 percent based on 20,000 test connections. This result was deemed to be highly satisfactory, as the exchange was very heavily loaded during peak periods.

Flower power and revolution

The cover of issue 3 in 1968 portrayed printed circuit cards in the transfer unit of Ericsson’s Stored Program Controlled (SPC) AKE exchange system. The photograph is from the automatic exchange installed in Tumba (a suburb of Stockholm). SPC exchanges were a milestone in the development of telephony, as they made it easier to trace faults, and thus reduce maintenace costs. The major concerns at the time were capacity and cost of memory.

Oil and energy crisesThe second issue of 1976

presented the AXE 10 switching system. The range of articles in this issue highlighted the shift in technology focus. Telecoms was becoming much more than wires, switches, exchanges and transmission. Hardware architecture and design were now being intimately combined with software structure to provide services, efficient traffic handling, management systems and scalability.




Software-defined networking: the service provider perspectiveAn architecture based on SDN techniques gives operators greater freedom to balance operational and business parameters, such as network resilience, service performance and QoE against opex and capex.

every year1. One of the root causes of network complexity lies in the tradi-tional way technology has developed. The design of network elements, such as routers and switches, has tradition-ally been closed; they tend to have their own management systems with verti-cally integrated forwarding and control components, and often use proprietary interfaces and features. The goal of network management is, however, to ensure that the entire network behaves as desired – an objective that is much more important than the capabilities of any single network element. In fact, implementing end-to-end networking is an important mission for most opera-tors, and having to configure individu-al network elements simply creates an unwanted overhead.

Network-wide programmability – the capability to change the behavior of the network as a whole – greatly simplifies the management of networks. And the purpose of SDN is exactly that: to be able to modify the behavior of entire net-works in a controlled manner.

The tradition of slow innovation in networking needs to be broken if net-works are to meet the increased demand for transport and processing capaci-ty. By integrating recent technological advances and introducing network-wide abstractions, SDN does just that. It is an evolutionary step in networking.

Telephony has undergone similar architectural transitions in the past. One such evolution took place when a clear separation between the functions

Ericsson’s approach to SDN goes beyond the data center addressing issues in the service-provider environment. In short Ericsson’s approach is Service Provider SDN. The concept aims to extend virtu-alization and OpenFlow – an emerging protocol for communication between the control and data planes in an SDN architecture – with three additional key enablers: integrated network control; orchestrated network and cloud man-agement; and service exposure.

There is no denying that networks are becoming increasingly complex. More and more functionality is being integrated into each network element, and more and more network elements are needed to support evolving service requirements – especially to support ris-ing capacity needs, which are doubling

AT T I L A TA K ACS, E L I SA BE L L AGA M BA, A N D JOE W I L K E

The traditional way of describing network architecture and how a network behaves is through the fixed designs and behaviors of its various elements. The concept of software-defined networking (SDN) describes networks and how they behave in a more flexible way – through software tools that describe network elements in terms of programmable network states.

The concept is based on split archi-tecture, which separates forwarding functions from control functions. This decoupling removes some of the com-plexity from network management, providing operators with greater flexi-bility to make changes.


API application programming interfaceARPU average revenue per user CLI command-line interface DPI deep packet inspectionGMPLS generalized multi-protocol label switching L2 Layer 2L3 Layer 3L2-L4 Layers 2-4M-MGW Mobile Media GatewayMME Mobility Management EntityMSC-S Mobile Switching Center ServerNAT Network Address TranslationNMS network management system

ONF Open Networking Foundation OSS/BSS operations and business support systemsPE provider edge devicePGW Packet Data Network GatewayRG residential gateway RWA routing and wavelength assignment SDN software-definednetworkingSGW Service GatewaySLA Service Level AgreementVHG virtual home gateway VoIP voice over IPWAN wide area network

20


New network abstraction layers

of the data plane (including SGW, PGW and M-MGW) and the control plane (including MME and MSC-S) was intro-duced. Now SDN has brought the con-cept of split architecture to networking.

As the business case proves, over the next two to five years, SDN technology will be deployed in networks worldwide. At the same time, the need to maintain traditional operational principles and ensure interoperability between SDN and more traditional networking com-ponents will remain. In the future, SDN will help operators to manage scale, reduce costs and create additional revenue streams.

StandardizationThe goal of the Open Networking Foundation (ONF), which was estab-lished in 2011, is to expedite the stan-dardization of the key SDN interfaces. Today, the work being conducted by ONF focuses on the continued evolu-tion of the OpenFlow protocol. ONF has recently established the Architecture and Framework Working Group, whose goal is to specify the overall architec-ture of SDN. The work carried out by this group will guide future standard-ization efforts based on strategic use cas-es, requirements for data centers and carrier networks, the main interfac-es, and their roles in the architecture. Ericsson is actively driving the work of this group, cooperating with other orga-nizations to promote the evolution of OpenFlow and support an open-source implementation of the most recent specifications.

Other standardization organizations, most notably the IETF, have recently begun to extend their specifications to support SDN principles. In IETF, the Interface to the Routing System (i2rs) WG and the recent activity of the Path Computation Element (PCE) WG will result in standardized ways to improve flexibility in changing how IP/MPLS net-works behave. This is achieved through the introduction of new interfaces to distributed protocols running in the network, and mechanisms to adapt net-work behavior dynamically to applica-tion requirements.

In addition to standardization organi-zations a multitude of active communi-ties and open-source initiatives, such as OpenStack, are getting involved in the

specification of various SDN tools, work-ing on maturing the networking aspect of virtualization.

Architectural visionSplit architecture – the decoupling of control functions from the physical devices they govern – is fundamental to the concept of SDN. In split-architecture networks, the process of forwarding in the data plane is separated from the con-troller that governs forwarding in the control plane. In this way, data-plane and control-plane functions can be designed, dimensioned and optimized separately, allowing capabilities from the underlying hardware to be exposed independently.

This ability to separate control and forwarding simplifies the development and deployment of new mechanisms, and network behavior becomes easi-er to manage, reprogram and extend. Deploying a split architecture, howev-er, does not remove the need for high-availability software and hardware components, as networks continue to meet stringent carrier requirements. However, the decoupling approach to architecture will help rationalize the network, making it easier to intro-duce new functions and capabilities. The ultimate goal of the SDN architec-ture is to allow services and applica-tions to issue requests to the network

dynamically, avoiding or reducing the need for human intervention to create new services. This, compared to today’s practices, will reduce the time to market of new services and applications.

The OpenFlow protocol2 is support-ing the separation of data and control planes and allowing the path of pack-ets through the network to be software determined. This protocol provides a simple abstraction view of networking equipment to the control layer. Split architecture makes virtualization of networking resources easier, and the control plane can provide virtual views of the network for different applications and tenants.

Ericsson has worked together with service providers to understand their needs for SDN both in terms of reduc-ing costs and creating new revenue opportunities. Based on these discus-sions, Ericsson has expanded the indus-try definition of SDN and customized it to fit the needs of operators.

Ericsson’s hybrid approach – Service Provider SDN – extends industry def-initions including virtualization and OpenFlow with three additional key enablers:

integrated network control;orchestrated network and cloud management; andservice exposure.

Service exposure

Northbound APIs to allow networks to responddynamically to application/service requirements

Integrated network control

Control of entire network from radio to edge tocore to data center for superior performance

Orchestrated network and cloud management

Unified legacy and advanced network, cloud management system and OSS/BSS to implementSDN in step-by-step upgrade

Service provider needs:

Service ProviderSDN

• Accelerated service innovation• Advanced public/hybrid enterprise and consumer cloud services• Improved QoE• Opex reduction through centralized management• Capex control

FIGURE 1 Service Provider SDN – components and promise

21


into this unified orchestration layer pro-vides the necessary support for applica-tions and tenants to trigger automatic changes in the network, ensuring opti-mal QoS and guaranteeing SLAs.

Unified and centralized orchestra-tion platforms greatly simplify the pro-cess of configuring, provisioning and managing complex service networks. Instead of having to tweak hundreds of distributed control nodes using fairly complex CLI programming, operations staff can use simple intuitive program-ming interfaces to quickly adjust net-work configurations and create new services. By accelerating the process of service innovation, Service Provider SDN will lead to increased market share and service ARPU, creating significant revenue growth and possibly reducing annual churn rates.

A high-level network architecture that supports the Service Provider SDN vision is illustrated in Figure   2. Service-provider networks will combine distributed control plane nodes (tradi-tional routers and appliances), and data-plane elements that are governed by centralized elements – SDN controllers – in the control plane. Consequently, to make service-provider networks pro-grammable, distributed and central-ized control-plane components must be exposed to a unified orchestration plat-form. In addition, the key elements of the orchestration platform and the con-trol plane need to be exposed to network and subscriber applications/services.

Application examplesThe major use cases or applications of

SDN in service-provider networks are summarized in Figure 3. Of these appli-cations, the data center was the first to make use of SDN. Ericsson’s approach to this is described in several articles pub-lished in Ericsson Business Review3 and in Ericsson Review4,5.

SDN can be applied in the aggrega-tion network to support sophisticated virtualization and to simplify the con-figuration and operation of this net-work segment. Ericsson has developed a proof-of-concept system in coopera-tion with Tier 1 operators to evaluate the applicability of SDN to aggregation net-works. This work has been carried out as part of the European Commission’s Seventh Framework Programme6.

Integrated network controlService providers will use SDN across the network from access, to edge to core and all the way into the data center. With integrated network control, oper-ators can use their network features, including QoS, edge functions and real- time activity indicators to deliver supe-rior user experience.

Orchestrated network and cloud managementService Provider SDN will integrate and unify legacy network management sys-tems with new control systems as well as with OSS/BSS. The platform for inte-grated orchestration supports end-to-end network solutions ranging from access over aggregation to edge func-tions as well as the data centers used to deliver telco and enterprise applications and services.

Service exposureNorthbound APIs expose the orches-tration platform to key network and subscriber applications and services. Together, the APIs and platforms allow application developers to maximize network capabilities without requiring intimate knowledge of their topology or functions.

By expanding the perspective of SDN to include these three elements, ser-vice providers can evolve their exist-ing network to the new architecture to improve the experience of their custom-ers. Implementing Service Provider SDN should remove the dumb pipe label giv-ing operators an advantage over com-petitors that do not own networks.

Network virtualizationOne of the benefits of Service Provider SDN, especially from a network- spanning perspective is network vir-tualization. Through virtualization, logical abstractions of a network can be exposed instead of a direct repre-sentation of the physical network. Virtualization allows logical topologies to be created, as well as providing a way to abstract hardware and software com-ponents from the underlying network elements, thereby separating control from forwarding capabilities and sup-porting the centralization of control.

Unified orchestration platforms sup-port network programming at the high-est layer as programming instructions flow through the control hierarchy – potentially all the way down to granular changes in flow paths at the forwarding plane level. Adding northbound APIs

Application

Control

Forwarding

Application

Control

Forwarding

(v) Control plane

APIs(such as OpenFlow)

Tenants

SDN controller

Forwarding Forwarding

Integratedsystems

Routers Physical

Forwarding elements

Ericsson NMS and cloud management system

Virtual

Software

Hardware

Application

FIGURE 2 Service Provider SDN – architectural vision

22



Ericsson and Telstra have jointly devel-oped a service-chaining prototype that leverages SDN technologies to enhance granularity and dynamicity of service creation. It also highlights how SDN can simplify network provisioning and improve resource utilization efficiency.

Packet-optical integration is a popular topic of debate, out of which several dif-ferent approaches are emerging. Split architecture provides a simple way to coordinate packet and optical network-ing; and so SDN, enhanced with features such as routing and wavelength assign-ment (RWA) and optical impairments management will be a natural fit for packet-optical integration.

Ericsson has started to develop solu-tions to virtualize the home gateway. Virtualization reduces the complexity of the home gateway by moving most of the sophisticated functions into the network and, as a result, operators can prolong the home gateway refreshment cycle, cut maintenance costs and accel-erate time to market for new services.

Virtualization of aggregation networksThe characteristics shared by aggrega-tion and mobile-backhaul networks are a large number of nodes and relative-ly static tunnels – that provide traffic grooming for many flows. These net-works are also known for their stringent requirements with regard to reliabil-ity and short recovery times. Besides L2 technologies IP and IP/MPLS is mak-ing an entrance as a generic backhaul solution. From an operational point of view, despite the availability of the dis-tributed control plane technology, this network segment is usually configured statically through a centralized man-agement system, with a point of touch to every network element. This makes the introduction of a centralized SDN controller straightforward for backhaul solutions.

A control element hosted on a telecom- grade server platform or on an edge router provides the operator with an interface that has the same look and feel as a single traditional router. The difference between operating an aggre-gation SDN network and a traditional network lies in the number of touch points required to provision and oper-ate the domain. In the case of SDN, only a few points are needed to control

the connectivity for the entire net-work. Consider, for example, an access/ aggregation domain with hundreds or even thousands of nodes running dis-tributed IGP routing protocols and the Label Distribution Protocol (LDP) to con-figure MPLS forwarding. In this case, SDN principles can be applied to simpli-fy and increase the scalability of provi-sioning and operating of such a network by pulling together the configuration of the whole network into just a few control points.

The control element treats the under-lying forwarding elements as remote line cards of the same system and, more specifically, controls their flow entries through the OpenFlow protocol. With this approach, any kind of connectivi-ty model is feasible regardless of wheth-er the forwarding node is L2 or L3 as, from a pure forwarding point of view, the same model is used in both flows. At the same time, network resilience at the transport level can be implement-ed by adding protection mechanisms to the data path. The SDN controller can pre-compute and pre-install back-up routes and then protection switch-ing is handled by the network elements for fast failover. Alternatively, the SDN controller can reroute around failures, in case multiple failures occur, or in

scenarios that have less stringent recov-ery requirements.

From the outside, the entire network segment appears to be one big PE rout-er and, for this reason, neighboring net-work elements of the SDN-controlled area cannot tell the difference (from a protocol point of view) between it and a traditional network. The network con-troller handles the interfacing process with legacy systems for connection set-up. Additional information on this point can be found in a presentation on the Virtual Network System7.

Dynamic service-chainingFor inline services, such as DPI, firewalls (FWs), and Network Address Translation (NAT), operators use different middle-boxes or appliances to manage subscrib-er traffic. Inline services can be hosted on dedicated physical hardware, or on virtual machines. Service chaining is required to route certain subscriber traffic through more than one such ser-vice. There are still no protocols or tools available for operators to perform flex-ible, dynamic traffic steering. Solutions currently available are either static or their flexibility is significantly limited by scalability inefficiencies.

Given the rate of traffic growth, con-tinued investment in capacity for

Virtualization ofaggregation network

Virtual homegateway

Policy-basedservice chaining

Packet and opticalintegration

Network supportfor cloud

Cloud/datacenterMobile

Residental

Business

FIGURE 3 Application examples

23


inline services needs to be managed carefully. Dynamic service-chaining can optimize the use of extensive high-touch services by selectively steering traffic through specific services or bypassing them completely which, in turn, can result in capex savings owing to the avoidance of over-dimensioning.

Greater control over traffic and the use of subscriber-based selection of inline services can lead to the cre-ation of new offerings and new ways to monetize networks. Dynamic service steering enables operators to offer sub-scribers access to products such as virus scanning, firewalls and content filters through an automatic selection and sub-scribe portal.

This concept of dynamic service chaining is built on SDN principles. Ericsson’s proof-of-concept system uses a logically centralized OpenFlow-based controller to manage both switches and middleboxes. As well as the traditional 5-tuple, service chains can be differen-tiated on subscriber behavior, applica-tion, and the required service. Service paths are unidirectional; that is, differ-ent service paths can be specified for upstream and downstream traffic.

Traffic steering has two phases. The first classifies incoming packets and assigns a service path to them based on predefined policies. Packets are then forwarded to the next service, based on the current position in their assigned

service path. No repeating classifica-tion is required; hence the solution is scalable.

The SDN controller sets up and recon-figures service chains flexibly to an extent that is not possible with today’s solutions. The dynamic reconfiguration of service chains needs a mechanism to handle notifications sent from middle-boxes to the controller. For example, the DPI engine notifies the controller that it has recognized a video flow. These notifications may be communicated using the extensibility features of the OpenFlow 1.x protocol.

Figure  4 summarizes service- chaining principles. The Virtual Network System (VNS) is a domain of the network where the control plane is centralized which excludes some of the traditional control agents. A simple API, such as OpenFlow, can be used to con-trol the forwarding functionality of the network, and the VNS can create north-bound interfaces and APIs to support creation of new features, such as service chaining, which allows traffic flows to be steered dynamically through servic-es or parts thereof by programming for-warding elements.

The services provided by the net-work may reside on devices located in different parts of the network, as well as within an edge router – for example, on the service cards of Ericsson’s Smart Services Router (SSR). Service chains are

programmed into the network based on a combination of information elements from the different layers (L2-L4 and pos-sibly higher). Based on operator policies, various services can be applied to traffic flows in the network.

For example, traffic may pass through DPI and FW functions, as illustrated by the red flow in Figure 4. However, once the type of the flow has been deter-mined by the DPI function, the opera-tor may decide to modify the services applied to it. For example, if the flow is an internet video stream, it may no lon-ger need to pass the FW service, reduc-ing load on it. Furthermore, after the service type has been detected, the sub-sequent packets of the same flow may no longer need to pass the DPI service either; hence the path of the flow can be updated – as indicated by the blue flow in Figure 4.

Packet-optical integrationThe increased programmability that SDN enables creates an opportunity to address the challenges presented by packet optical networking. SDN can simplify multi-layer coordination and optimize resource allocation at each lay-er by redirecting traffic (such as VoIP, video and web) based on the specific requirements of the traffic and the best serving layer.

Instead of a layered set of separated media coordinated in a static manner, SDN could transform the packet-optical infrastructure to be more fluid, with a unified recovery approach and an allo-cation scheme based on real-time link utilization and traffic composition. The ONF still has some work to do to adapt OpenFlow to cope with optical constraints.

To speed up packet-optical inte-gration, a hybrid architecture can be deployed where OpenFlow drives the packet domain, and the optical domain remains under the control of GMPLS. This approach utilizes the extensive optical capabilities of GMPLS and, therefore, instead working to extend OpenFlow with optical capabilities, it allows us to focus on the actual integra-tion of optical and packet domains and applications that utilize the flexibility of a unified SDN controller.

FIGURE 4 Service-chaining principles

Virtual Network Systemdomain

FWDPI NAT FW

DPISSR

SDNswitch

Flow before

Flow after

SDNswitch

SDNswitch

24



Home gateway controlThe concept of the virtual home gateway (VHG) introduces a new home-network architecture primarily driven by con-sideration to improve service delivery and management. The target architec-ture emerges by applying SDN capabil-ities between the residential gateway (RG) and the edge network – moving most of the gateway’s functionalities into an embedded execution environ-ment. Virtualizing the RG significant-ly reduces its complexity and provides the operator with greater granularity in remote-control management, which can be extended to every home device and appliance. As a result, operators can reduce their investments significantly by prolonging the RG refreshment cycle, cutting maintenance costs and acceler-ating time to market for new services.

The VHG concept allows operators to offer seamless and secure remote profile instantiation stretching the boundaries of a home network without compromis-ing security. The concept provides the tools to configure and reconfigure mid-dleboxes dynamically, so that commu-nication between devices attached to different home networks can be estab-lished, and/or provide specific connec-tivity requirements for a third-party service provider – between, for exam-ple, a utility company and a particular device. By embedding SDN capabilities, Ericsson’s concept enables operators to offer personalized applications to sub-scribers each with its own specific chain

of management policies and/or services.The target architecture places an oper-ator-controlled bridge at the custom-er’s premises instead of a complex router, while the L3-L7 functionalities are migrated to the IP edge or into the operator cloud. Using SDN technology between the IP edge and the switch in this way offers the operator fine-grained control for dynamic configuration of the switch.

ConclusionWith its beginnings in data-center tech-nology, SDN technology has developed to the point where it can offer signif-icant opportunities to service provid-ers. To maximize the potential benefits and deliver superior user experience, SDN needs to be implemented outside the sphere of the data center across the entire network. This can be achieved through enabling network programma-bility based on open APIs.

Service Provider SDN will help operators to scale networks and take advantage of new revenue- generating possibilities. The broader Service Provider SDN vision goes beyond lever-aging split architecture to include sev-eral software components that can be combined to create a powerful end-to-end orchestration platform for WANs and distributed cloud data centers. Over time, this comprehensive software-based orchestration platform will be able to treat the overall operator net-work as a single programmable entity.

The authors would like to thank Diego Caviglia, Andreas Fasbender, Howard Green, Wassim Haddad, AlvarodeJodra,IgnacioMás,Don McCullough and Catherine Truchan for their contributions to this article.

Acknowledgements

Joe Wilke

is head of Development Unit IP & Broadband Technology Aachen. He currently leads the SDN

execution program and holds a degree in electrical engineering from the University of Aachen, Germany and a degree in engineering and business from the University of Hagen, Germany.

Elisa Bellagamba

is a portfolio strategy manager in Product Area IP & Broadband, where she has been leading SDN

related activities since their very beginning. She holds an M.Sc. cum laude in computer science engineering from Pisa University, Italy.

Attila Takacs

is a research manager in the Packet Technologies Research Area of Ericsson Research. He has been the

technical lead of research projects on Software Defined Networks (SDN), OpenFlow, GMPLS, Traffic Engineering, PCE, IP/MPLS, Ethernet, and OAM for transport networks. He is also an active contributor to standardization; in particular he has worked for ONF, IETF and IEEE. He holds more than 30 international patent applications; granted and in progress. He holds an M.Sc. in computer science and a post-graduate degree in banking informatics, both from the Budapest University of Technology and Economics, in Hungary. He has an MBA from the CEU Business School, in Budapest.

1. Ericsson, November 2012, Mobility Report, On the pulse of the Networked Society, available at: http://www.ericsson.com/ericsson-mobility-report

2. Open Networking Foundation, available at: https://www.opennetworking.org/ 3. Ericsson, November 2012, Ericsson Business Review, The premium cloud:

how operators can offer more, available at: http://www.ericsson.com/news/121105-ebr-the-premium-cloud-how-operators-can-offer-more_244159017_c

4. Ericsson Review, December 2012, Deploying telecom-grade products in the cloud, available at: http://www.ericsson.com/res/thecompany/docs/publications/ericsson_review/2012/er-telecom-grade-cloud.pdf

5. Ericsson Review, December 2012, Enabling the network-embedded cloud, available at: http://www.ericsson.com/res/thecompany/docs/publications/ericsson_review/2012/er-network-enabled-cloud.pdf

6. European Commission, Seventh Framework Programme, Split Architecture Carrier Grade Networks, available at: http://www.fp7-sparc.eu/

7. ElisaBellagamba:VirtualNetworkSystem,MPLS&EthernetWorldCongress,Paris,February2012,available at: http://www.slideshare.net/EricssonSlides/e-bellagamba-mewc12-pa8

References

25


HSPA evolution: for future mobile-broadband needsAs HSPA continues to evolve, addressing the needs of changing user behavior, new techniques develop and become standardized. These techniques provide network operators with the flexibility, capacity and coverage needed to carry voice and data into the future.

capacity – to handle growing smartphone traffic cost-efficiently;flexibility – to manage the wide range of traffic patterns efficiently; andcoverage – to ensure good voice and app user experience everywhere.

App coverageFor smartphone applications, like social networking and video streaming, to function correctly, access to the data net-work and a network that can deliver a defined minimum level of performance is needed. The relationship between the performance requirements (in terms of data speed and response time) of an application and the actual performance delivered by the network for that user at their location at a given time deter-mines how well the user perceives the performance of the application.

The term app coverage denotes the level of network performance needed to provide subscribers with a satisfactory user experience for a given application. In the past, the task of dimension-ing networks was simpler, as calcula-tions were based on delivering target levels of voice coverage and providing a minimum data rate. Today’s appli-cations, however, have widely vary-ing performance requirements. As a result, dimensioning a network has become a more dynamic process and one that needs to take these varying performance requirements into con-sideration, for apps that are currently popular with subscribers.

FootprintIllustrated in Figure 2, at the end of 2012, 55 percent of the world’s popu-lation was covered by WCDMA/HSPA, a figure that is set to rise to 85 percent

voice-centric feature phones. For less than USD 100, consumers can purchase highly capable WCDMA/HSPA-enabled smartphones with dual-core processors and dual-band operation that support data rates up to 14.4Mbps. This price-to- sophistication ratio has turned the smartphone into an affordable mass-market product, and has accelerated the increase in smartphone subscriptions – estimated to rise from 1.2 billion at the end of 2012 to 4.5 billion by 20181.

Ericsson ConsumerLab studied a group of people to assess how they per-ceived network quality and what issues they encountered when using their smartphones. The study identified two key factors that are essential to the per-ceived value of a smartphone: a fast and reliable connection to the data network, and good coverage2.

These findings highlight an impor-tant goal for operators: to provide all network users with high-speed data ser-vices and good-quality voice services everywhere. This can be achieved by securing:

N I K L A S JOH A NSSON, L I N DA BRUS, E R I K L A R SSON, BI L LY HOGA N A N D PET E R VON W RYCZA

BOX A Terms and abbreviationsCELL_FACH Cell forward access channelCPC Continuous Packet ConnectivityDPCH Dedicated Physical ChannelEUL Enhanced UplinkHS-DSCH High-Speed Downlink Shared ChannelHSDPA High-speed Downlink Packet AccessHSPA High-speed Packet AccessHSUPA High-speed Uplink Packet AccessLPN low-power node

M2M machine-to-machineMBB mobile broadbandMIMO multiple-input multiple-outputROT rise-over-thermalSRB Signaling Radio BearerUL uplinkURA_PCH UTRAN registration area paging channelUTRAN Universal Terrestrial Radio Access NetworkWCDMA Wideband Code Division Multiple Access

Mobile broadband (MBB), providing high-speed internet access from more or less anywhere, is becoming a reality for an increasing proportion of the world’s population. There are several factors fuelling the need for high-performance MBB networks, not the least, the growing number of mobile internet connections. As Figure 1 illustrates, global mobile subscriptions (excluding M2M) are predicted to grow to 9.1 billion by the end of 2018. Nearly 80 percent of mobile subscriptions will be MBB ones1, indicating that MBB will be the primary service for most operators in the coming years.

Impact of affordable smartphonesTo a large extent, the rapid growth of MBB can be attributed to the widespread availability of low-cost MBB-capable smartphones, which are replacing

26


Smarter networks

by the end of 20181. Today, many devel-oped markets are nearing the 100 per-cent population coverage mark3. This widespread deployment, together with support for the broadest range of devic-es, makes WCDMA/HSPA the primary radio-access technology to handle the bulk of MBB and smartphone traffic for years to come.

Since its initial release, the 3GPP WCDMA standard has evolved, and continues to develop. Today, WCDMA/HSPA is a best-in-class voice solution with exceptional voice accessibility and retainability. It offers high call retention as well as being an excellent access tech-nology for MBB, as it delivers high data rates and high cell-edge throughput – all of which enable good user experience across the entire network.

The continued evolution of WCDMA/HSPA in Releases 11 and 12 includes sev-eral key features that aim to increase network flexibility and capacity to meet growing smartphone traffic and secure voice and app coverage.

Evolution of traffic patternsApplications have varying demands and behaviors when it comes to when and how much data they transmit. Some apps transmit a large amount of data continuously for substantial periods of time and some transmit small pack-ets at intervals that can range from a few seconds to minutes or even longer. Applications have varying demands, typically sending lots of data in bursts, interspersed with periods of inactivity when they send little or no data at all.

Rapid handling of individual user requests, enabled by high instantaneous data rates, improves overall network performance as control-channel over-head is reduced and capacity for other traffic becomes available sooner. So, if a network can fulfill requests speedily, all users will experience the benefits of reduced latency and faster round-trip times.

Web browsing on a smartphone is a classic example of a bursty application, both for uplink and downlink commu-nication. When a smartphone requests the components of a web page from the network (in the uplink) they are trans-ferred in bursts (in the downlink), and the device acknowledges receipt of the content (in the uplink). As a result,

uplink and downlink performance becomes tightly connected and there-fore better uplink performance has a positive effect on downlink data rates as well as overall system throughput.

For web browsing, the instantaneous downlink speed for mobile users needs to be much higher on average than the uplink speed. However, the number of services requiring higher data rates in

the uplink, such as video calling and cloud synching of smartphone data, is on the rise.

As user behavior changes, traffic- volume patterns also change, and mea-surements show it is becoming more common for uplink levels to be on par with downlink levels, and in some cases even outweigh the downlink traf-fic. Consequently, continuing to

FIGURE 1 Mobile and MBB subscriptions (2009-2018)1

Mobile subscriptionsMobile broadband

2009 2010 2011 2012 2013 2014 2015 2016 2017 20180

2,000

1,000

3,000

4,000

6,000

5,000

8,000

7,000

9,000

10,000

Subscriptions/lines (million)

FIGURE 2 Population coverage by technology (2012-2018)

100

80

60

40

20

2012 2018 2012 2018 2012 20180

% p

op

ulat

ion

cove

rag

e

(Source: Ericsson1)

>85%>90%

~55%

>85%

~10%

~60%

GSM/EDGE WCDMA/HSPA LTE

27


develop data rates to secure uplink-heavy services is key to improving over-all user performance.

High performance networksThe standard approach used to create a high-performance network with wide coverage and high capacity is to first improve the macro layer, then densify it by deploying additional macro base sta-tions, and finally add low power nodes (LPNs) in strategic places, such as traf-fic hotspots, that can offload the macro network.

Each step addresses specific perfor-mance targets and applies to different population densities, from urban to rural – as illustrated in Figure 3. The evolution of WCDMA/HSPA includes a number of features that target macro layer improvement and how deploy-ments where LPNs have been added can be enhanced.

Improving the uplinkFeatures in the 3GPP specification have recently achieved substantial improve-ment of uplink capabilities. Features such as uplink multi-carrier, higher-order modulation with MIMO, EUL in CELL_FACH state, and Continuous Packet Connectivity (CPC) have multi-plied the peak rate (up to 34Mbps per carrier in Release 11) and increased the number of simultaneous users a net-work can support almost fivefold.

Given the high uplink capabilities already supported by the standard, the next development (Release 12) will enable and extend the use of these capa-bilities to as many network users as possible.

The maximum allowed uplink inter-ference level in a cell, also known as maximum rise-over-thermal (ROT), is a highly important quantifier in WCDMA networks. This is because the maxi-mum allowed interference level has a direct impact on the peak data rates that the cell can deliver.

Typically, macro cells are dimen-sioned with an average ROT of around 7dB, which enables UL data rates of 5.7Mbps (supported by most commer-cial smartphones), and secures voice and data coverage for cell-edge users. High data rates, such as 11Mbps (avail-able since release 7) and 34Mbps (avail-able since release 11) require ROT levels

FIGURE 3 Where to improve, densify and add

ImproveDensifyAdd

ImproveDensify

Improve

Denseurban

Urban Suburban Rural

Area traffic density

FIGURE 4 Relationship between maximum interference and peak rate

UL ROT

Rate

Y

X

Y = Maximum interference handled by the network

X = Maximum uplink data rate that can be achieved

Legend

28


Smarter networks

greater than 10dB and 20dB respective-ly. Figure 4 illustrates the relationship between ROT and peak data rate.

The maximum uplink interference level permissible is determined by a number of factors including the density of the network, the capability of the net-work to handle interference (for exam-ple with advanced techniques such as Interference Suppression), and the capa-bilities of the devices in the network, including both smartphones and legacy feature phones.

The Lean Carrier solution, introduced in Release 12, is an additional capability that helps operators meet the needs of high-data-rate users. This multi-carrier solution is built on the Release 9 HSUPA dual-carrier one that is currently being implemented in commercial smart-phones. The dual-carrier solution allows two carriers, primary and secondary, to be assigned to a user. By doing this, the traffic generated by the user can be allocated in a flexible way between the two carriers, while at the same time doubling the maximum peak rate achievable.

The Lean Carrier solution optimizes the secondary carrier for fast and flexi-ble handling of multiple high-data-rate users, through more efficient granting and lower cost per bit. The solution is designed to support multiple bursty data users in a cell transmitting at the highest peak rates without causing any uplink interference among themselves or to legacy users. To maximize ener-gy efficiency, the Lean Carrier solution should cost nothing in system or termi-nal resources on the secondary carrier until the user starts to send data.

Lean Carrier can be flexibly deployed according to the needs of the network. For example, the maximum ROT on a user’s secondary (lean) carrier can be configured to support any available uplink peak data rate, while the max-imum ROT on a user’s primary carrier can be configured to secure cell-edge coverage for signaling, random access and legacy (voice) users.

Rate adaptation is another technolo-gy under study that results in increased network capacity for some common traffic scenarios, such as areas where subscribers are a mix of high and low-rate users or areas where there are only high-rate users. High uplink data rates

require more power. Maintaining a fixed data rate at the desired quality target in an environment where interference levels vary greatly can result in large fluctuations in received power. To avoid such fluc-tuations, the concept of rate adapta-tion can be applied. High-rate users are assigned with a fixed received-power budget, and as interference levels change, bit rates are adapted to maintain the desired quality tar-get, while not exceeding the allowed power budget. In short, as illustrated in Figure 5, the bit rate is adapted to received power, and not the power to the rate.

Limiting fluctuations in received power for high-rate users is good for overall system capacity because these high-rate users can transmit more efficiently, and other users in the system, including low-rate ones such as voice users, consume less power when power levels are stable and predictable.

Maintaining a device in connected mode for as long as possible is anoth-er technique that can be used to improve performance of the uplink.

Smartphone users want to be able to rapidly access the network from a state of inactivity. Maintaining a device in a connected-mode state, such as CELL_FACH or URA_PCH, for as long as possible is one way of

achieving this – access to the network from these states is much faster than from the IDLE state. In recent releases, connected mode has been made more efficient from a battery and resource point of view through the introduction of features such as CPC, fractional DPCH and SRB on HS-DSCH. As a consequence it is now feasible to maintain inactive devices in these states for longer.

As the number of smartphone users increases, networks need flexi-ble mechanisms to maintain high sys-tem throughput, even during periods of extremely heavy load. Allowing the network to control the number of con-currently active users, as well as the number of random accesses, is one such mechanism.

Improvements that enable high throughput under heavy load, and allow users to benefit from lower latency in connected mode, while enabling ser-vice-differentiated admission decisions and control over the number of simul-taneous users, have been proposed for Release 12.

Expanding voice and app coverageGood coverage is crucial for positive smartphone user experience and cus-tomer loyalty2, which for operators translates into securing voice cover-age and delivering data-service cover-age that meets the needs of current and future apps.

FIGURE 5 Rate adaptation results in predictable interference levels

Baseline: Fixed ratevariable power

Received power

DATA

Control

DATA

Control

Time

Rate adaptation: Fixed received power and variable rate

29


Uplink transmit diversity was intro-duced in Release 11. This feature sup-ports terminals with two antennas to increase the reliability and coverage of uplink transmissions and decrease overall interference in the system. It works by allowing the device to use both antennas for transmission in an effi-cient way using beamforming. Figure 6 illustrates how the radio transmission becomes focused in a given direction, resulting in a reduction in interfer-ence between the device and other nodes, and improving overall system performance.

An additional mode within uplink transmit diversity is antenna selection. Here, the antenna with the best radio propagation conditions is chosen for transmission. This is useful, for exam-ple, when one antenna is obstructed by the user’s hand. Uplink transmit diversity increases the coverage of all uplink traffic for voice calls and data transmissions.

With Release 11, multi-flow HSDPA transmissions are supported. This allows two separate nodes to transmit to the same terminal, improving per-formance for users at the cell edge and resulting in better app coverage.

In Release 12, simultaneous app data and voice call transmissions will become more efficient, and the time it takes to switch the transmission time interval from 10ms to 2ms is consid-erably shorter. These improvements increase both voice and app coverage.

Enhancing small-cell deploymentsThe addition of small cells through deploying LPNs in a macro network – resulting in a heterogeneous network – is a strategic way to improve capacity, data rates and coverage in urban areas. Typically, the deployment of LPNs is ben-eficial in hotspots where data usage is heavy, to bridge coverage holes created by complex radio environments, and in some specific deployments such as in-building solutions.

Figure 7 shows the performance gains in a heterogeneous-network deployment (which is described in more detail in Box B). Offloading to small cells has several benefits: it provides increased capacity for handling smart-phone traffic, and it results in enhanced app coverage.

FIGURE 7 System-level gains – for scenario described in Box B

1W5W

0

50

Average Cell edge

100

150

200

250

300

User throughput gain (percent)

 BOX B   The systemThe scenario shown in Figure 7 is for bursty traffic. Four LPNs have been added to each macro base station in the network, and 50 percent of the users are located in traffic hotspots. The transmission power for the macro base station was 20W, and 1W and 5W LPNs were deployed.

LPNs were deployed randomly and no LPN range expansion was used. Gains are given relative to a macro-only deployment. Offloading was 32 percent for 1W LPNs and 41 percent for 5W LPNs, where offloading is a measure of the percentage of traffic served by the LPN.

FIGURE 6 Release 11 uplink transmit diversity beamforming

There are several ways to improve coverage for voice and data. One way is to use lower frequency bands – when compared to 2GHz bands, considerable coverage improvement can be achieved by refarming the 900MHz spectrum

from GSM, for example. Voice coverage can be significantly extended with lower-rate speech codecs, whereas four-way receiver diversity and advanced antennas can improve coverage for both voice and data.

30


Smarter networks

1. EricssonMobilityReport,June2013,availableat: http://www.ericsson.com/res/docs/2013/ericsson-mobility-report-june-2013.pdf

2. EricssonConsumerLabreport,January2013,Smartphoneusageexperience–the importance of network quality and its impact on user satisfaction, available at: http://www.ericsson.com/news/130115-ericsson-consumerlab-report-network-quality-is-central-to-positive-smartphone-user-experiences-and-customer-loyalty_244129229_c

3. International Communications Market Report 2011, Ofcom, available at: http://stakeholders.ofcom.org.uk/binaries/research/cmr/cmr11/icmr/ICMR2011.pdf

References

FIGURE 8 LPN deployment scenarios

LPN LPN

LPN

Macro

LPNs deployed as separate cells on the same carrier

RNC

LPN LPN

LPN

Macro

LPNs deployed as part of a combined cell on the same carrier

RNC

31


To maximize spectrum usage, the tra-ditional macro base stations and LPNs share the same frequency, either with separate or shared cell identities. These deployments, illustrated in Figure 8, are referred to as separate cell and com-bined cell.

It is possible to operate both separate and combined-cell deployments based on functionality already implemented in the 3GPP standard, and such deploy-ments have been shown to provide sub-stantial performance benefits over macro-only deployments.

Today, combined cells tend to be deployed in specific scenarios, such as railroad, highway and in-building envi-ronments. Separate-cell deployments, on the other hand, are more generic and provide a capacity increase in more common scenarios.

In 3GPP Release 12, small-cell range expansion techniques and control chan-nel improvements are being introduced to enable further offloading of the macro network. Mobility performance enhancements for users moving at high speeds through small cell deployments are also being investigated by 3GPP.

When a macro cell in a combined-cell deployment is complemented with additional LPNs close to users, the data rate and network capacity is improved. By allowing the network to reuse the same spreading codes in different parts of the combined cell, the cell’s capac-ity can be further increased – a tech-nique being studied in Release 12. And as there is no fundamental uplink/ downlink imbalance in a combined cell, mobility signaling is robust, signaling load is reduced, and network manage-ment is simplified.

In summary, heterogeneous net-works are essential for handling grow-ing smartphone traffic because they support flexible deployment strategies, increase the capacity of a given HSPA network, and extend voice and app cov-erage. The improvements standardized in Release 12 will further enhance these properties.

ConclusionsWCDMA/HSPA will be the main tech-nology providing MBB for many years to come. Operators want WCDMA/HSPA networks that can guarantee excellent user experience throughout the whole

network coverage area for all types of current and future mobile devices. The prerequisites for networks are:

capacity – to handle growing smartphone traffic cost-efficiently;flexibility – to manage the wide range of traffic patterns efficiently; andcoverage – to ensure good voice and app user experience everywhere.

HSPA evolution, through the capabili-ties already available in 3GPP and those under study in 3GPP Release 12, aims to fulfill these prerequisites. There are several ways to improve voice and app

coverage. Enhancements to the uplink improve the ability to quickly and effi-ciently serve bursty traffic – improv-ing user experience and increasing smartphone capacity. Small-cell improvements will increase network capacity for smartphone traffic and fur-ther improve voice and app coverage.

With all of these enhancements, WCDMA/HSPA, already the dominant MBB and best-in-class voice technology, has a strong evolution path to meet the future demands presented by the growth of MBB and highly capable smartphones globally.

Niklas Johansson

is a senior researcher at Ericsson Research. He joined Ericsson after receiving his M.Sc. in

engineering physics and B.Sc. in business studies from Uppsala University in 2008. Since joining Ericsson, he has been involved in developing advanced receiver algorithms and multi-antenna transmission concepts. Currently, he is project manager for the Ericsson Research project that is developing concepts and features for 3GPP Release 12.

Peter von Wrycza

is a senior researcher at Ericsson Research, where he works with the development and

standardization of HSPA. He received an M.Sc. (summa cum laude) in electrical engineering from the Royal Institute of Technology (KTH), Stockholm, Sweden, in 2005, and was an electrical engineering graduate student at Stanford University, Stanford, CA, in 2003-2005. In 2010, he received a Ph.D. in telecommunications from KTH.

Erik Larsson

joined Ericsson in 2005. Since then has held various positions at Ericsson Research,

working with baseband algorithm design and concept development for HSPA. Today, he is a system engineer in the Technical Management group in the Product Development Unit WCDMA and Multi-Standard RAN and works with concept development and standardization of HSPA. He holds an M.Sc. in engineering physics (1999) and a Ph.D. in signal processing (2004), both from Uppsala University, Sweden.

Billy Hogan

joined Ericsson in 1995 and works in the Technical Management group in the Product Development Unit WCDMA and Multi-

Standard RAN. He is a senior specialist in the area of enhanced uplink for HSPA. He works with the system design and performance of EUL features and algorithms in the RAN product, and with the strategic evolution of EUL to meet future needs. He is currently team leader of the EUL Enhancements team for 3GPP release 12. He holds a B.E. in electronic engineering from the National University of Ireland, Galway, and an M.Eng in electronic engineering from Dublin City University, Ireland.

Linda Brus

joined Ericsson in 2008. Since then, she has been working with system simulations, performance

evaluations, and developing algorithms for WCDMA RAN. Today, she is a system engineer in the Technical Management group in the Product Development Unit WCDMA and Multi-Standard RAN, working with concept development for the RAN product and HSPA evolution. She holds a Ph.D. in electrical engineering, specializing in automatic control (2008) from Uppsala University, Sweden.

32


Smarter networks

Next generation video compressionMPEG and ITU have recently approved a new video-compression standard known as High Efficiency Video Coding (HEVC), or H.265, that is set to provide double the capacity of today’s leading standards1.

the internet. It is the preferred format for Blu-ray discs, telepresence streams and, most notably, HDTV.

Now imagine a codec that is twice as efficient as H.264. This was the tar-get set by MPEG and ITU in 2010, when they embarked on a joint standardiza-tion effort that three years later deliv-ered HEVC/H.2654,5.

The new codec offers a much more efficient level of compression than its predecessor H.264, and is particu-larly suited to higher-resolution video streams, where bandwidth savings with HEVC are around 50 percent. In simple terms, HEVC enables a network to deliv-er twice the number of TV channels. Compared with MPEG-2, HEVC can pro-vide up to four times the capacity on the same network.

Like most standards, the MPEG and ITU video codecs have been developed in a collaborative fashion involving many stakeholders – manufacturers, operators, broadcasters, vendors and academics. Ericsson has been an active participant in video standardization for more than 15 years, and was closely involved in HEVC.

Throughout the development of the standard, Ericsson has led several of the core experiments, chaired ad-hoc working groups and contributed signif-icantly to the development of the tech-nology behind the codec. Our greatest expertise lies in the areas of the deblock-ing filter6  and in reference picture management7.

Concepts that create efficiencyOne of the primary target areas for HEVC compression is high- resolution video, such as HD and UHD. The statis-tical characteristics of these types

devices – from phones and tablets to PCs and home-entertainment systems. As HD shifts from luxury to commodity, it will soon be challenged by UHD, which offers resolutions up to 16 times greater.

Making standardsMost video viewed by subscribers today has been digitized and reduced in size through the application of a com-pression standard. The more popular include the H.26x series from ITU and the MPEG-x series from ISO/IEC. First published in 1994, the MPEG-2 stan-dard, also known as H.262, played a crucial role in the launch of digital-TV services as it enabled the compression of TV streams to fit the spectrum avail-able. This is also the standard used to compress movies onto a DVD.

The H.264 standard (also known as MPEG-4 AVC), published in 2003, has provided the best compression efficien-cy to date, and is currently the most widely used video-compression codec. It has been successfully incorporated into most mobile devices, and is the best way to reduce the size of video carried over

PE R F RÖJ DH, A N DR EY NOR K I N A N D R ICK A R D SJÖBE RG


AVC advanced video coding CABAC context-adaptive binary arithmetic coderCTU coding-tree unitCU coding unitfps frames per secondHD highdefinition;oftenrefersto 1280 x 720 or 1920 x 1080 pixelsHEVC HighEfficiencyVideoCodingIEC International Electrotechnical Commission

ISO International Organization for Standardization ITU International Telecommunication Union MPEG Moving Picture Experts GroupOTT over-the-topSAO sample adaptive offsetUHD ultrahighdefinition:often refers to 3840 x 2160 (4K) or 7680 x 4320 (8K) pixelsWPP wavefront parallel processing

Requiring only half the bitrate of its predecessor, the new standard will significantly reduce the need for bandwidth and expensive, limited spectrum. HEVC will enable new video services to be launched, and the first applications that are likely to appear will be for mobile devices and OTT applications, followed by TV – and in particular ultra HD television (UHDTV).

State-of-the-art video compression can reduce the size of raw video by a fac-tor of about 100 without any noticeable reduction in visual quality. Estimates indicate that compressed real-time vid-eo accounts for more than 50 percent of current network traffic2, and this fig-ure is set to rise to 90 percent within a few years3.

New services, devices and changing viewing patterns are among the fac-tors contributing to this growth, as is increased viewing of traditional TV and video-streaming services, such as Netflix, YouTube and Hulu, on a range of

33


Same bandwidth, double the data

of video streams tend to be different from lower- resolution content: frame sizes are larger, and frame rates and per-ceived quality are higher – imposing tough requirements on compression efficiency, as well as on the computa-tional complexity of the encoding and decoding processes.

As the architectures of smartphone s and tablets go multi-core, the ability to take advantage of parallel process-ing is key when it comes to the efficient compression of high-resolution content. All of these points have been taken into

FIGURE 1 Simplified HEVC encoder diagram

Input signal

Intra-prediction

In-loopfilters

Entropycoding

Decodedpicture buffer

Motioncompensation

Motionestimation

Transform coefficients

Motion vectors

T Q

Q-1

T-1

FIGURE 2 Example of the coding-tree unit structure in HEVC

CTU structureCTU

CU

CU

CU

consideration during the development of the new standard.

The hybrid block-based coding used by the new codec is the same as the one used in earlier video-coding stan-dards. To encode content, video frames are divided into blocks that are coded individually by applying prediction – based either on neighboring blocks in the same picture (intra prediction) or from previously coded pictures (motion estimation/compensation).

The difference between the pre-dicted result and original video data is

subsequently coded by applying block transforms and quantization. In this way, a block can be represented by just a few non-zero coefficients. Quantized transform coefficients, motion vectors, prediction directions, block modes and other types of information are encod-ed with lossless entropy coding. Hybrid block-based coding is illustrated in Figure  1.

To ensure the highest level of com-pression efficiency, and support for par-allel processing, some parts of HEVC have been significantly modified com-pared with previous generations of hybrid block-based codecs. For most of the previous MPEG-x and H.26x codecs, the largest entity that could be inde-pendently encoded was a macroblock (16 × 16 pixels). For HEVC, the picture is split into coding-tree units (CTUs) with a maximum size of 64 × 64 pix-els. Every CTU is the root of a quadtree, which can be further divided into leaf-level coding units (CUs), as illustrated in Figure  2. The CTUs are coded in ras-ter scan order, and each unit can itself contain a quadtree structure. Each CU contains one or more prediction parti-tions that are predicted independent-ly of each other. A CU is also associated with a transform quadtree that com-presses the prediction residual and has a structure similar to that of a CTU – as shown in Figure  2.

Partitions for motion prediction can form square or rectangular shapes, which is also the case with earlier stan-dards. HEVC also supports something called asymmetric motion partition-ing, which can split the CU into predic-tion units of unequal width or height, as illustrated in Figure  3.

The size of the prediction blocks in HEVC can therefore vary from 4 × 4 sam-ples up to 64 × 64, while transform sizes vary from 4 × 4 to 32 × 32 samples. Large prediction blocks and transform sizes are the most efficient way to encode large smooth areas, whereas smaller prediction blocks and transforms can be used to achieve precision in areas that contain finer detail.

The HEVC specification covers more intra-prediction modes than H.264, including a planar mode to approx-imate a surface from neighboring pixels, a flat mode and 33 angular pre-diction modes. Motion-compensated

34



The concept behind WPP is to re-ini-tialize CABAC at the beginning of each line of CTUs.

To facilitate CABAC adaptation to the content of the video frame, the coder is initialized once the statistics from the decoding of the second CTU in the pre-vious row are available.

Re-initialization of the coder at the start of each row makes it possible to begin decoding a row before the pro-cessing of the preceding row has been completed. Thus, as shown in the exam-ple in Figure  4, several rows can be decoded in parallel in several threads

with a delay of two CTUs between two consecutive rows.

The Tiles tool can be used for paral-lel encoding and decoding, and works by dividing a picture into rectangular areas (tiles) – as shown in Figure  5 – where each tile consists of an integer number of CTUs. The CTUs are pro-cessed in a raster scan order within each tile, and the tiles themselves are processed in the same way. Prediction based on neighboring tiles is disabled, and so the processing of each tile is inde-pendent. In-loop filters, however, can operate over tile boundaries. And

prediction for luma transform blocks is performed with up to quarter-pixel precision, whereas motion compensa-tion for color components is performed with one-eighth-of-a-pixel precision. Interpolation for fractional pixel posi-tions uses 8-tap filters for luma blocks and 4-tap filters for color.

In HEVC there is a single entropy coder for low-level data. This is the con-text-adaptive binary arithmetic coder (CABAC), which is similar to the one used in H.264, but modified to facilitate parallel processing. Higher-level infor-mation, such as sequence parameters, is encoded with variable-length or fixed-length encoding.

HEVC defines two in-loop filters: a deblocking filter and a sample adaptive offset (SAO) filter. The latter is applied to the output of the deblocking filter, and increases the quality of reference pic-tures by applying transmitted offsets to samples that fulfill certain criteria. In-loop filters improve the subjective quality of reconstructed video as well as compression efficiency. Deblocking filtering in HEVC is less complex than that of H.264, as it is constrained to an 8 × 8 block grid. This constraint, togeth-er with filtering decisions and opera-tions that are non-overlapping between two boundaries, simplifies multi-core processing.

Parallel processingTo make the most of the increasingly widespread use of multi-core proces-sors, plus the ever-growing number of cores used in consumer-class proces-sors, significant attention was paid to the parallelization characteristics of video encoding and decoding when designing HEVC. As it is computation-ally more complex than its predecessor, maximizing parallelization has been a key factor in making HEVC an effi-cient real-time encoding and decoding solution.

Several HEVC tools have been designed for easy parallelization. The deblocking filter can be applied to 8 × 8 pixel blocks separately, and transform-coefficient-coding contexts for several coefficient positions can be processed in parallel. Tiles and wavefront paral-lel processing (WPP) are among several HEVC tools that can provide high-level parallelism.

Asymmetric motion partitions are shown in the bottom row. Only square partitions are allowed for intra prediction

FIGURE 3 Possible motion prediction partitions in HEVC

CTU

FIGURE 4 Multi-thread decoding with wavefronts. Gray areas indicate CTUs that have already been decoded

35


as deblocking and SAO can be par-allelized, filtering can be performed independently inside each tile, and tile boundaries can be processed by in-loop filters in a final pass.

The HEVC standard therefore enables both high- and low-level parallelization, which can provide significant benefits for multi-thread encoding and decod-ing of video such as 4K and 8K that has a higher resolution than HD.

Performance and complexityThe improved coding efficiency of HEVC does however come with a price tag: increased computational complex-ity. Compared with its predecessor, HEVC is 50 -100 percent more complex for decoding and up to 400 percent more complex when it comes to encoding. While these comparisons are based on preliminary tests, they do give an indi-cation of the new codec’s computation-al complexity.

Real-time implementations of HEVC demonstrate that decoding of full HD (1080p) at 50 or 60fps is possible on fast desktop and laptop computers, running on a single core. Performance increas-es with multiple core implementations (hardware acceleration), so that a mod-ern smartphone is capable of 1080p decoding at 25 or 30fps8.

Applications The new standard is a general one suit-able for the compression of all kinds of video. The focus for the first version is consumer applications and for this, three profiles have been defined: Main, Main 10 and Main Still Picture.

Main is an all-purpose profile with a depth of 8 bits per pixel, supporting 4:2:0 – the most common uncompressed video format used by consumer devices from mobile phones to HDTVs. Main 10 extends the bit depth to 10 bits per pixel, which is well suited to consumer appli-cations, such as UHDTV, where very high quality is critical. The increased bit depth can compress wide dynamic range video without creating band-ing artifacts, which sometimes occurs with 8 bits. The third profile, Main Still Picture, used for still images, is a subset of Main and carries a single still picture at a depth of 8 bits per pixel.

The initial deployments of HEVC released in 2013 will be for mobiles and OTT applications. Software imple-mentations capable of decoding HEVC without hardware acceleration can easily be downloaded to smartphones, tablets and PCs, enabling mobile TV, streaming and download services on existing devices. To this end, in August 2012, Ericsson announced SVP 55009, the world’s first HEVC real-time video

encoder for live-TV delivery to mobile devices. However, as it is better to per-form encoding on hardware and as HEVC is computationally more demand-ing than previous standards, it may be some time before video telephony based on this standard enters mobile platforms, whereas encoding on PCs is already feasible.

Set-top boxes with new decoders will become available soon, enabling con-tent broadcast via satellite, cable or ter-restrially to take advantage of HEVC. The new standard plays a key role in the provision of UHDTV, and as prices drop and displays become affordable, the number of services utilizing such high resolutions is expected to rise within a few years. Flat-panel displays for HDTV have been on the market for almost a decade, so this may be a good time for consumers to start upgrading to UHDTV.

What’s coming The finalized version of HEVC targets most consumer devices and services. However, for more specialized applica-tions, such as 3D, content production or heterogeneous devices and networks, some additions to HEVC may prove use-ful. With this in mind, MPEG and ITU are working together on a number of ideas, including support for stereo and multi-view (glasses-free) 3D video, an extension that encodes multiple views by rearranging picture buffers and reference picture lists. A first drop is expected in January 2014, with a more advanced version that will support joint encoding of texture and depth informa-tion coming in the early part of 2015.

Scalability is a key attribute of any codec, as it enables trimming of video streams to suit different network con-ditions and receiver capabilities; scal-able extensions to HEVC are planned for July 2014. Range extensions, which support several color formats as well as increased bit depths, are another area currently under development.

In addition to these extensions, fur-ther improvements are expected to take place inside the current HEVC frame-work, such as more efficient encoding and decoding (both software and hard-ware). It is likely that the full potential of HEVC will take some time to unfold, as encoding algorithms develop and the

FIGURE 5 Example of the way an image can be divided into tiles

CTU

CTU

CTU

Column boundaries

Row boundaries

tile 1 tile 2 tile 3



36



challenge posed by the optimization of encoders and decoders in multi-core architectures is overcome.

In short, HEVC or H.265 is twice as efficient as its 10-year-old predecessor, H.264. The improved efficiency that this codec brings will help to ease traffic load in networks and enable the creation of new and advanced video-based services.

The codec supports parallel process-ing and even though it is more complex from a decoding perspective, tests have shown that it is suitable for adoption in mobile services. Compression of mobile video streams and OTT content are the most likely initial candidates for applica-tion of the codec, and within a few years it will undoubtedly bring UHDTV into our homes.

Per Fröjdh

is director of media standardization at Ericsson and former head of visual technology at

Ericsson Research. He holds an M.Sc. in engineering physics and a Ph.D. in theoretical physics from Chalmers University in Gothenburg, Sweden. Part of his Ph.D. work was carried out on scholarship at Imperial College London, UK. Following postdoctoral appointments in the US, and Denmark, he held the position of professor of theoretical physics at Stockholm University, Sweden. He joined Ericsson in 2000 as manager of video research and standardization. He has contributed to MPEG and ITU work on H.264 and HEVC, served on the advisory committee for the W3C, and has been the editor of 15 standards on streaming, file formats, and multimedia telephony in MPEG, ITU, 3GPP and IETF.

Andrey Norkin

is a senior researcher at Ericsson Research, Kista, Sweden. He holds an M.Sc. in computer

engineering from Ural State Technical University, Yekaterinburg, Russia and a Ph.D. in signal processing from the Tampere University of Technology, in Finland. He has worked at Ericsson Research since 2008, contributing to HEVC standardization through technical proposals and activities, including the coordination of a core experiment on deblocking filtering, chairing break-out groups and subjective quality tests for the Joint Collaborative Team on Video Coding (JCT-VC). He has also been active in the 3D video standardization for JCT-3V. He is currently the project manager of the 3D VISION project at Ericsson Research, working on 3D video systems, and algorithms, as well as on parts of the standardization.

Rickard Sjöberg

is a senior specialist in video coding in the Multimedia Technologies department at Ericsson

Research, Kista, Sweden. With an M.Sc. in computer science from the KTH Royal Institute of Technology, Stockholm, Sweden, he has been working with Ericsson since 1997 and has worked in various areas related to video coding, in both research and product development. In parallel, he has been an active contributor in the video-coding standardization community, with more than 100 proposals relating to the H.264 and HEVC video-coding standards. He is currently working as the technical leader of Ericsson’s 2D video-coding research, including HEVC and its scalable extensions. His research interests include video compression and real-time multimedia processing and coding.

1. ITU,January2013,pressrelease,Newvideocodec to ease pressure on global networks, availableat:http://www.itu.int/net/pressoffice/press_releases/2013/01.aspx#.UWKhxBnLfGc

2. Ericsson, November 2012, Mobility Report, available at: http://www.ericsson.com/ericsson-mobility-report

3. Fierce Broadband Wireless, 2013, Ericsson CEO: 90%ofnetworktrafficwillbevideo,availableat:http://www.fiercebroadbandwireless.com/story/ericsson-ceo-90-network-traffic-will-be-video/2013-02-25

4. ITU-T Recommendation H.265 | ISO/IEC 23008-2:HighEfficiencyVideoCoding,availableat:http://www.itu.int/ITU-T/recommendations/ rec.aspx?rec=11885

5. IEEE, December 2012, Overview of the High EfficiencyVideoCoding(HEVC)Standard,available at: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6316136

6. IEEE,December2012,HEVCDeblockingFilter,available at: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6324414

7. IEEE,December2012,OverviewofHEVCHigh-Level Syntax and Reference Picture Management, available at: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6324417

8. IEEE, DOCOMO Innovations, December 2012, HEVCComplexityandImplementationAnalysis,available at: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6317152

9. Ericsson, 2012, Ericsson announces world’s firstHEVCencoderforliveTVdeliverytomobiledevices, available at: http://www.ericsson.com/news/120822_ericsson_announces_worlds_first_hevc_encoder_for_live_tv_delivery_to_mobile_devices_244159018_c

References

37


Next generation OSS/BSS architectureBreaking down the silos of operations and business support systems (OSS/BSS) to form an integrated, cross-functional platform that can take a product from conception to execution in a simplified and consistent manner will cut time to market from months and years to weeks.

evolve to cater for factors such as big data, personalization and virtualiza-tion? The simple answer is through configurability. To create a high level of flexibility, the evolution of OSS/BSS needs to be configuration driven, with an architecture based on components.

The impact of big data Information is a critical resource. Good information is a key asset – one that can be traded, and one that is critical for optimizing operations. As volumes rise, the rate of creation increases, and a wid-er variety of data that is both structured and unstructured floods into OSS/BSS, access to storage needs to be effortless. In this way, tasks and optimization pro-cesses can maximize the use of existing infrastructure and keep data duplica-tion to a minimum.

Data management needs to be secure and controllable, ensuring that the

The challenges of evolutionBy exposing the functionality and infor-mation held in their networks, opera-tors have the opportunity to create innovative and ever more complex val-ue chains that include developers, OTT players and subscribers. In these new value chains, the flow of information and control shifts from unidirection-al to multidirectional, and participants can be consumers of services and infor-mation as well as being producers of them.

New business models for network evo-lution are based on providing anything as a service (XaaS) – including IaaS, PaaS, SaaS and NaaS – and when using this model, it is not just value chains that become more complex; the life cycles of products and services also become more diversified.

How then, as business models advance, should OSS/BSS requirements

JA N F R I M A N, L A R S A NGE L I N, E DVA R D DR A K E A N D M U N I SH AGA RWA L


BO business objectBPMN Business Process Model and NotationBSS business support systemsCEP complex event processing

CLI command-line interface(E)SP (enterprise) service busETL extract, transform, loadeTOM enhanced Telecom Operations MapGUI graphical user interfaceIA information architecture IaaS infrastructure as a serviceIM information model JEE JavaEnterprise Edition

LC life cycle LDAP Lightweight Directory Access ProtocolM2M machine-to-machineNaaS network as a serviceNFV network functions virtualizationOLAP online analytical processingOLTP online transaction processingOS operating systemOSGi OSGi Alliance (formerly Open Services Gateway Initiative)OSS operations support systemsOTT over-the-topPaaS platform as a servicePO purchase order

RAM random access memorySaaS software as a serviceSBVR SemanticsofBusinessVocabulary and Business RulesSDN software-definednetworkingSID shared information/data modelSLA Service Level AgreementSQL Structured Query LanguageTCO total cost of ownershipTMF TeleManagement ForumUI user interfaceVM virtual machineXaaS anything as a service

The systems that keep networks running and profitable are in the direct line of fire when it comes to implementing change. So, as the world moves toward global connectivity, as smartphones cause a shift in user behavior, and as subscribers demand more personalized products and even greater control, the functions of OSS/BSS – such as planning, configuration, fulfillment, charging, billing and analytics – need to be integrated.

A consolidated architecture is a typical computer-science approach for bringing together the functions of different sys-tems. By adopting such a consolidated architecture for OSS/BSS, operators will be able to maintain control over costs while implementing network changes effectively.

38


The merger of two giants

systems accessing information do not jeopardize data integrity and subscrib-ers can feel confident that their infor-mation is protected.

The impact of subscriber needsPersonalized services and superior user experience are key capabilities for business success and building loy-alty. Subscribers want to be in control, and feel that their operator provides them with reasonably priced services that meet their individual needs, over a network that delivers near real-time response times. The ability to create and test services in a flexible way with short time to market will help operators meet changing user demands.

The impact of M2MAs the number of connected devices gets closer to 50 billion, the need for automated and autonomous behavior in processes such as configuration and provisioning is becoming more signifi-cant. Being able to remotely configure, provision and update millions of devic-es without impacting the network sup-ports scaling while maintaining control over opex.

The impact of virtualizationAs a result of virtualization, operators, partners and even subscribers (in the future) can create instances of their ser-vices and networks on demand. So, as networks continue to move into the cloud, and SDN and NFV technologies become more widespread, the number of entities managed by OSS/BSS will rise by several orders of magnitude.

So, to help operators remain compet-itive, next generation OSS/BSS need to fully address the challenges created by certain aspects of network evolution, including virtualization, big data, M2M and personalization.

Making good use of technologyOne way to address these challenges is to make good use of advancing tech-nology, particularly when it comes to OSS/BSS implementation architecture. And it’s not just about using technol-ogy development in a smart way; it’s also about understanding the potential of a given technology. So, when a new concept results in a significant break-through, the services and products that

can be created as a result should be read-ily definable.

Capitalizing on increased flexibility and agility made possible by new tech-nologies (such as virtualization and SDN) needs to be coordinated through a management function, which puts new demands on OSS/BSS architecture.

The evolution of virtualizationThe demands created by increasing vir-tualization of data centers, not just in terms of computational capacity, but also in terms of storage and networking capabilities, are:

virtualization of the OSS/BSS, and running these systems in the cloud;

FIGURE 1 Separating planes in SDN architecture

Business application(such as charging and billing)

Businessapplicationplane

Controlplane

Dataplane

SDNcontroller

ND ND ND ND ND ND ND ND

ND ND ND ND ND ND ND

SDN architectureND = network device

SDNcontroller

SDNcontroller

SDNcontroller

Business application(such as CRM)

Business application(such as network inventory)

Business application(such as VPN service)

FIGURE 2 Abstraction of a typical OSS/BSS deployment

Gap/disconnect

Enterprise’s business

Present BSS/OSS

TraditionalBSS/OSSfunction

TraditionalBSS/OSSfunction Traditional

BSS/OSSfunction










39


FIGURE 3 Extracting hard-coded business logic

Gap/disconnect


Today Tomorrow


Enterprise’s virtualized BSS/OSS

Present BSS/OSS

Traditional BSS/OSSapp

Business rules

Business processes

Business events

Information

.......

Integration

Storage

Analytics



BSS/OSSfunction

BSS/OSSfunction

Sharedinformation

model

Commonstorage

Commonanalytics

Commonintegration

Businessrule

Businessprocess

Ericsson BSS/OSSapplication logic

Businessevent

Businessobject

Enterpriseinformation

model




virtualization, where virtual machines (VMs) share the resources of a single hardware host, is evolving so network infrastructure is becoming more effi-cient. For example, the failover capabil-ities of the hypervisor can place agents on the host in a similar way to tradi-tional failover clusters, and can moni-tor not only VM health, but application and OS health as well. Such features are prerequisites of an efficient virtual environment. However, application architecture may have to take these features into account, as in some cases they cause the responsibility to perform certain tasks (such as data recovery) to shift between the application and the infrastructure.

Service provider SDNSDN separates the data plane (the for-warding plane or infrastructure layer) from the control plane, which in turn, is separated from the business applica-tion plane. As shown in Figure  1, vari-ous business applications communicate with SDN controllers, providing a virtu-alized – possibly hierarchical – view of the underlying data plane.

Generally speaking, the management

requirements for SDN and non-SDN architectures are similar, if not the same. For example, both require inven-tory, ordering and fault management. However, SDN presents a new set of tech-nical issues related to resource manage-ment, which brings into question the current partitioning and structure of OSS/BSS architectures.

Specifically, SDN can result in hori-zontally abstracted virtualized software layers that have limited vertical vision through the hierarchy from the busi-ness applications to network devices. So, at the same time as the abstraction offered by SDN makes it easier to expose the capabilities of the network, it creates additional challenges for the OSS/BSS architecture.

OSS always needs to have the capa-bility to map the virtual view of the network to the underlying implemen-tation. Sometimes, the SDN controller handles domain-specific OSS/BSS func-tionality by hiding parts of the com-plexity of the control and data planes; and as a result, only a subset of infor-mation will be propagated to the busi-ness application plane. Sometimes, the underlying layers are not even vis-ible – such as when a third party owns them. In these cases, SLAs can be used to map the virtual view to the underly-ing implementation.

The evolution to SDN architecture and virtualization causes the number of entities managed by OSS/BSS compo-nents to rise, which in turn impacts the way they are managed. For example, a more extensive history of each entity is required, because the semi- static envi-ronment used to locate a device using its IP address no longer exists. With SDN, network topology becomes total-ly dynamic. History data is essential for analysis and management of the net-work, as this information puts network events into context.

Hybrid flexibilityModern database design is evolving toward the use of hybrid architectures. This approach allows a wider range of solutions and applications to be created with a single consistent implementa-tion and one logical data store.

Hybrid disk/in-memory databases use in-memory technologies to achieve the performance and low latency levels

management of cloud-based OSS applications such as service assurance; andmanagement of cloud-based BSS applications such as IaaS and PaaS.

It may, however, not always be benefi-cial to run certain network elements on generic IaaS resources. For example, information stored in a database may be better provided in the form of a service to subscribers in an IaaS environment, rather than as virtually deployed ten-ants. The general rule is that anything provided as a service, which is imple-mented by a piece of software running in a generic IaaS environment, has a reduced level of control and efficiency. Due to the extra layers created by run-ning software in a generic environ-ment, the drawbacks of this approach must be weighed carefully against the benefits of increased flexibility and bet-ter (shared) use of physical resources.

For next generation OSS/BSS, the focus should be placed on implementing flexibility in an efficient way together with automation and orchestration of resource allocation.

The hypervisor approach to

40



of an in-memory solution, while still using disk for data persistency. The hybrid approach allows more data to be stored on disk than can fit into mem-ory; as such, the disk is not a mirror of the in-memory content. This approach is similar to caching disk content, while providing the performance that comes from a true in-memory design – which cannot be achieved by caching disk con-tent alone.

Hybrid SQL/NoSQL (sometimes referred to as NewSQL) solutions are SQL-capable databases that are built using a NoSQL implementation to attain the scalability and distribution that these architectures afford, while still providing support for SQL. However, such hybrid solutions can be limited by their lack of support for partition-wise joins and subsequent lack of support for ad hoc queries – although there are exceptions to this. Hybrid OLTP/OLAP solutions aim to merge the typical characteristics of transactional OLTP workloads and OLAP-based workloads (related to analytics) into a single imple-mentation. To build such a structure typically requires that both database architectures be considered from the outset. Even if such solutions exist, it is difficult to build this type of hybrid from the starting point of an OLTP- or OLAP-optimized architecture.

Big data Modern data centers are designed so that increasingly large amounts of mem-ory with low levels of latency are being placed ever closer to computational resources. This greatly increases the level of real-time processing that can be achieved as well as the volumes of data that can be processed. Achieving these processing levels is not simply a matter of the speed at which operations can be carried out; it is also about creating new capabilities. The developments being made in big-data processing have a sig-nificant impact on how next generation OSS/BSS architecture can be designed.

Fast data – the velocity attribute of big data – is the ability to make real-time decisions from large amounts of data (stored or not) with low latency and fast processing capabilities. Fast data sup-ports the creation of filtering and corre-lation policies that are based on – and can also be adjusted to – near real-time input.

Another big-data concept com-bines the in-memory/disk hybrid with the OLTP/OLAP (row/column) hybrid to achieve a single solution that can address both OLTP and demanding analytics workloads. Coupled with the huge amounts of memory that modern servers can provide, this approach removes the need for a separate analytics database.

The business logicWhen OSS/BSS are deployed, they bring business and technical stake-holders together and allow them to focus on the design and implemen-tation of their unique business. The functionality provided by OSS/BSS must support the necessary user-friendly tools to implement and develop business logic. The deeper and more flexible this support is, the more business opportunities can be

explored, and the more profitable an enterprise can be.

Figure 2 shows an abstraction of a typical OSS/BSS deployment. Current implementations tend to be multiven-dor, with multiple systems performing similar tasks. Organic growth has led to a lack of coordination and as a result, sig-nificant time and effort is spent on inte-gration, time to market is long, and TCO tends to be high. Quite often there a sig-nificant gap between daily business and the systems used to support it.

To transform such a complex archi-tecture into a more business-agile system requires some evolution. As con-solidation is fairly straightforward, this tends to be the first step. However, to succeed, it requires systems to be mod-ular, to be able to share data, to use data from other sources, and to have a spe-cific role. Another approach is to reduce the silo nature of OSS/BSS and instead be more flexible using business logic in

FIGURE 4 An order handling process. The notation shown here is conceptual; processes are modeled in BPMN1 2.0 and rules in SBVR2

Purchaseorder

Receivepurchase order

Sendpurchase order

delivered

Sendpurchase order

reception acknowledgement

Orderhandling

Customerhandling

Productoffering

Createcustomer

order

Purchaseorder

received

Checkcustomer

order

Activateservices

Activate customer order

Activateresources

Customerorder

Customerorder

Customerorder

Activateservice

Activateresources

Activatebilling

Archivecustomer

order

Activatebilling

Archivecustomer

order

External

Requiredentities

Event andprocess

Created BOand LC

Applicationtasks

41


from the underlying systems, and to structure functionality according to design and life-cycle flow. To achieve this and build an abstract and virtual view of the business successfully, a common, shared and semantically rich information model (IM) and a defined set of relationships are essential.

a configurable manner. Either way, the best evolution approach is in system architecture design.

Architecture proposalAs visualized in Figure 3, the Ericsson approach to next generation OSS/BSS is to extract hard-coded business logic

The information model is the fun-damental component of Ericsson’s approach to next generation OSS/BSS.

Building blocksWhen the hard-coded business logic is extracted the following building blocks are created:

actors and roles – such as companies, functions, individuals and customers, suppliers and service providers;services and functions – such as sales and contracting;processes – such as TMF eTOM;business objects – such as products, orders, contracts and accounts; andrules – such as pricing, prioritization and product/service termination.

These building blocks can then be used to form processes, as Figure 4 illus-trates. To achieve the full degree of flexibility, the building blocks for the complete life cycle (from definition to termination) are needed.

Conception to executionFigure 5 illustrates design chain func-tionality. This process is relied upon to take a new business idea from concep-tion to execution.

Business logic is defined, designed and implemented in what Ericsson refers to as the enterprise business stu-dio. The studio comprises a set of inte-grated workbenches that have access to all building blocks and informa-tion elements, and provide feedback to the process owner as business logic is implemented.

In the Ericsson model, business log-ic (such as verification, commissioning and decommissioning, supervision, migration and optimization) is trans-ferred to a management function for implementation and application. The management function is also responsi-ble for transferring business logic to the proper execution engines. For example, pricing rules are used in execution by the rating engine, contracting engine and the sales portal engine. Given the potentially massive spread of pricing rules, getting it right at this stage of development is key.

Information architectureGenerally speaking, an enterprise defines the information it needs to

FIGURE 5 Business logic from conception to execution

Feedback

Businesslogicdecision

C-levelmanagement

Businessarchitects

Enterpriseand ITarchitects

Businesslogicdesign

Businesslogicimplement

Businesslogicmanagement

Executablebusinesslogic

Strategy Design Implement Deploy Operate

Enterprise’sBSS/OSSmanagement

Enterprise’sBSS/OSSengines

Enterprise’s business studio

FIGURE 6 Information models within an enterprise

Enterprise productionenvironments

Mediation/transformation

Enterprise data model

Otherenterprise systems PartnerDeployed OSS/BSS architecture

Transformation

Ericsson applicationlevel information model

Other

Other

Otherapplication level

informationmodel

Enterprise’s applicationlevel information model

Enterprisebusiness concepts model

Enterprise OSS/BSS architecture reference model

42



operate, and the OSS/BSS manage this information, based on a range of busi-ness models. The spread of information across any given enterprise is extensive, and can span many different functional areas from marketing, ordering, strate-gy and HR, to production and finance. Information models used in OSS/BSS include:

the enterprise vocabulary and concepts at the business level – for example, an enterprise might refer to a voice product using its marketing name, such as Family and Friends; the canonical concepts at the application level – which might refer to the Family and Friends product as family-group; and the multivendor concepts at the application level – where the concept of Family and Friends has different names at the business level and the canonical level.

As operators continue to differentiate and offer ever more complex products and services, the requirements on infor-mation change. Information is no longer just mission critical; it is also enterprise critical, and changes constantly as busi-ness needs evolve.

The shift to next-generation OSS/BSS changes the way enterprise- critical information needs to be handled, creat-ing a number of system requirements:

information and applications need to be separated;the entire life cycle – from definition to termination – needs to be modeled;information needs to be shared among all enterprise users;master data needs to be determined, and even multiple masters need to be supported to align with different enterprise functions; and information needs to be characterized in terms of size, throughput, quality, reliability and redundancy across the board, for all functions and applications – one instance that can be used by all.

As Figure 6 shows, information held in an enterprise can be generated by many sources – both internal and exter-nal – and can be classified according to its properties and type. For example, information can be static, structured, event-driven or transactional.

The best way to meet the new

requirements on information – driven by the need to differentiate – is to design OSS/BSS in a way that is independent of functional applications with a centrally managed information architecture (IA) that has a common and shared informa-tion model.

The key characteristics of this archi-tecture are:

integrated information across functions;information offered as a service – facilitating high-level abstraction and avoiding the need to understand low-level data constructions; andinformation published in catalogs – formalizing IaaS and enabling use by process owners.

Due to the complexity and widespread nature of information models, a mod-ularized information architecture is needed – one that can be configured to meet the varying needs of enterprises and used in multivendor scenarios with varying information life cycles.

Modular layersA modular IA can be implemented by categorizing information into a matrix.

The first step is to categorize informa-tion into (horizontal) layers, where each layer is populated by a number of enti-ties. Subsequently, these entities can be combined (vertically) into a solution.

As Figure 7 shows, the data vault is placed at the bottom of the architec-ture hierarchy. The most efficient type of storage can be chosen from a number of components, including disk, RAM or virtual resources.

A level up from the data layer is the engine layer, which comprises a set of components that provide access to the information storage. SQL, LDAP, NoSQL or HBase are examples of technologies used in this layer, and each group of data selects the technology that best matches the access requirements for that data.

The grid layer, above the engine, is where the information model becomes accessible. Here, the IM is divided into a set of responsibility areas, such as enter-prise catalog with product, service and resource specifications or inventory data. These areas deploy components that expose information as a service and protect the underlying data,

FIGURE 7 Information architecture implementation

Data access and grid services

Master data

Reference and ID

Communication buses

Metadatafunctionality

Informationmodel

Non-real-time API

Customerand partner

management

EnterpriseOSS/BSSinformationmanagement

Applicationlayer

Data grid framework

Data engine framework

Data vault

Enterprisecatalog

Eventmanagement

Analyticsmanagement

NonEricsson

apps

Bi-d

irection

al integ

ration

and

transfo

rmatio

nservices

Data access and grid services Real-time API

Non-real-time API Data engine services Real-time API

PublishedIM

Information accessTransactionsupport

(E)SBdistribution

Referenceand global ID

Analytics and CEPengines

Storage schemasengine

Storage

Information serviceregistry

Integrationmaster/slave

Data access

IM

Data transform

Grid and vault

External data sourcesconnection

Data engineservices

Analytics and SEPengines

Data persistance

Enterprise

43


Management functions such as def-inition, registration, discovery, usage, archiving and decommissioning are shown on the left of the information architecture. As information is no lon-ger hard-coded in each application, but shared among applications at all stages of the life cycle, the manage-ment function is vital for ensuring data consistency.

Deployment stack To serve next generation OSS/BSS, a state-of-the-art deployment stack is required. A functional view of such a stack is illustrated in Figure 8. The deployment stack should provide a con-sistent user experience for all processes – from business configuration and sys-tem provisioning, to operations, admin-istration and management. It should support applications deployed on a vari-ety of different infrastructures includ-ing cloud, and virtualized and bare metal hardware. The deployment stack should provide a means of efficient inte-gration among applications, and enable service exposure in a uniform way.

ensuring that it is consistent and available for all authorized applications.

The applications that consume and produce information sit on the top of the architecture. They set the require-ments on the grid layer so that infor-mation is made available with the right characteristics and accessibility accord-ing to the needs of the given applica-tion. Applications can of course use local caching, but the data vault allows infor-mation to persist and be made available throughout the system.

The right-hand side of Figure 7 shows the set of functions that support the flow of information in and out of the model. Typically, this part of the OSS/BSS inter-faces to data owned by legacy or oth-er systems, and allows information to be accessed from the outside. Grid com-ponents are responsible for interacting with external data sources, and expos-ing access to information residing in the model. Typical functions include trans-formation, protocol adaptations, han-dling of services or data streams and identity mapping, which are also used by applications in the application layer.

FIGURE 8 A functional view of the deployment stack for next generation OSS/BSS

Presentation

Business level

Information tier Applications

Application services In-service performance Licensing User management Coordination service

Configuration managementLogging

JEE/OSGi

Linux File system

Availability support

Load balancer

Backup and restoreSoftware management

Performance management

Graphical user interface Command line interface Machine-to-machine interface

Fault managementOperations andmanagement

Middleware

IP

OS

Virtualization

Hardware

Clouddeployment

By making use of the common ser-vices provided by a deployment stack that supports scale-out architecture and meets the needs of big data, application development should become more effi-cient. The stack should integrate easi-ly with existing enterprise systems – a capability that becomes more signifi-cant as OSS/BSS are developed and used in multiple scenarios around the world and deployed in an IaaS manner.

HardwareUsing existing hardware infrastruc-ture for OSS/BSS deployment is the best option for operators as it consolidates the use of hardware, and supports rapid reconfigurability and scalability. Linux is an attractive OS, as it is community driven and supports the decoupling of software and hardware elements.

MiddlewareA common and well-composed middle-ware provides: a consistent environ-ment, effective management, ease of integration, greater availability, bet-ter scaling, load balancing, simplified installation, upgrade and deployment, and improved backup capabilities. Such an environment can be provid-ed by either an OSGi container or a JEE application server pre-integrated with availability management, soft-ware management and backup/restore capabilities.

Operations and managementThis function provides common man-agement services for configuration, log-ging, configuration management, and fault and performance management.

Application servicesThis layer provides common functions for OSS/BSS applications, such as:

service performance – which monitors and reports uptime; licensing – which enables provisioning, monitoring, control and reporting of licenses; user management – which provides authentication and authorization; and coordination service – which provides inter-application coordination in a distributed environment that supports changing license requirements created by business models such as pay-as-you-grow.

44



Presentation layerThe presentation layer provides support for GUI, CLI and M2M interfaces for OSS/BSS applications. A common GUI frame-work, together with single sign-on for the entire stack, is key to providing a consistent user experience.

By exposing interfaces to other appli-cations in a uniform way, the amount of application-application integration required is reduced significantly.

Proposed architectureThe architecture of next generation OSS/BSS is illustrated in Figure 9. At the business level, the proposal sup-ports service agility in all processes from conception to retirement and all relevant phases – including planning, deployment, customer on-boarding and assurance.

The proposed architecture compris-es a set of application functions, which implement tasks such as enterprise catalog, charging, billing, order man-agement, experience and assurance. Application functions are implement-ed through a set of components that can be configured and assembled so they form a complete solution that can also be integrated with existing systems. The set of common application func-tions, including correlation and event handling, support the specific OSS/BSS application functions.

The information architecture sepa-rates the information model so infor-mation is matched to the application functions, supporting modularity and enabling integration in the overall infor-mation model. The information model is based on a shared information/data model (SID), with extensions embrac-ing more standard industry informa-tion models.

To support effective implementation, all components should be prepared for cloud deployment.

FIGURE 9 Next generation OSS/BSS architecture

OSS/BSS

BSS/OSSapplicationfunctions

App levelmanagement

App servicegovernance

App IM andmaster datagovernance

ApplicationSW compliance

andgovernance

CEP and ETLmanagement

Ap

plicatio

n inform

ation m

od

e

Business stud

io

Custo

mer and

partner

manag

ement

Custo

mer and

partner

interaction

Enterp

risecatalo

g

Ord

erm

anagem

ent

Reso

urcem

anagem

ent

Deb

tm

anagem

ent

Revenue

manag

ement

Exp

erience andassurance

Commonapplicationfunctions

Application level information tier

Rule

framew

ork

Co

ntentm

anagem

ent

Wo

rkflowfram

ewo

rk

Do

cument

form

at

Analytics

Co

rrelation

CE

P

Event

handling

Decisio

nsup

po

rt

Co

mm

unication

channels

Presentation layer

Business level

Application level

Application services

Operations and management

Middleware

IP

OS

Virtualization

Hardware

Common operations and management GUI

Software managementGUI/CLI

Business studio UI

45


Jan Friman

is an expert in the area of user and service management at Business Unit Support Systems (BUSS). He has held

various positions in the area of OSS/BSS at Ericsson for 16 years, including R&D, system management and strategic product management. He is chief architect for information architecture at BUSS and holds an M.Sc. in computer science from Linköping Institute of Technology, Sweden.

Lars Angelin

is an expert in the technology area multimedia management at BUSS. Lars has more

than 28 years of work experience in the areas of concept development, architecture and strategies within the telco and education industries. Lars joined Ericsson in 1996 as a research engineer, and in 2003 he moved to a position as concept developer for telco-near applications, initiating and driving activities, most of them related to M2M or the OSS/BSS area. He holds an M.Sc. in engineering physics and a Tech. Licentiate in tele-traffic theory from Lund Institute of Technology, Sweden.

Edvard Drake

is an expert in the area of hardware and software platform technologies, and is chief architect for

implementation architecture at BUSS. He has 20 years’ experience at Ericsson, ranging from AXE-10 exchanges to today’s commercial and open source technology innovation. He holds a B.Sc. in software engineering from Umeå University, Sweden.

Munish Agarwal

is a senior specialist in multimedia architecture and chief implementation architect for OSS/BSS. He

has been at Ericsson since 2004, working in the OSS/BSS area. He is currently driving the BOAT implementation architecture and is the product owner for Next Generation Execution Environment. He holds a B.Tech. in material science from the Indian Institute of Technology, Kharagpur, India.

1. Business Process Modeling Notation, available at: http://www.bpmn.org

2. SemanticsofBusinessVocabularyand Rules, available at: http://www.omg.org/spec/SBVR

References

46



One of the most significant technical evolutions in the history of telecoms is that of the mobile phone. In 1990, Ericsson received its first order to supply a GSM network, which brought the company’s switching and radio expertise together. At the same time, the internet began its worldwide expansion, as did the liberalization of the telecom market.

In 1993, Sweden became the first European country to deregulate its telecom market. While already a fact in the US, deregulation has led to increased competition and a greater need for innovation to deliver customer benefits.

In 1994, after almost a decade of research, Ericsson released the Bluetooth technology standard, which allows devices to exchange information wirelessly. In the spirit of open stan-dards, its control has been handed over to the Bluetooth Special Interest Group (SIG) – a non-profit organization.

That same year, Ericsson Review carried an article about intelligent network architecture for the Japanese 2G digi-tal cellular standard Personal Digital Cellular (PDC). This enhanced network architecture was based on the principle of strictly separating network-oriented services from the mobile subscriber-specific services, allowing for the rapid develop-ment of new services in response to demand.

At the turn of the 21st century, Ericsson addressed the chal-lenges of VoIP. The surpassing of voice by data – a fact that didn’t become reality until the end of 2009 – was clear, and the primary challenge was how to port voice services to the new packet-based platform, while maintaining the same lev-el of quality.

The first decade of the new century was dominated by LTE and the desire to evolve network architecture to support tech-nology evolution, improve spectral efficiency, be more flexible and ultimately support new services and superior user expe-rience. In February 2007, Ericsson demonstrated LTE with bit rates of up to 144Mbps, and theoretical peak rates of 1.2Gbps were demonstrated in 2010. In 2009, Ericsson delivered the first commercial LTE network.

And last but not least, the dramatic impact on networks created by smartphones, tablets and other mobile devices, the increased demand for mobile broadband and soaring data traffic have ramped up the opportunities in telecoms. Ericsson today holds a unique position in that we can provide network equipment, solutions, services, and management tools that encompass the full spectrum of needs.

Looking forward, one of our visions for the Networked Society is that everything that can benefit from a connection will be connected. Connectivity is the starting point for new ways of innovating, collaborating and socializing.



47

Re:view

Ultra modernismIn 1984, an entire issue of

Ericsson Review was dedicated to fiber, providing a report on fiber optic activities at Ericsson. The articles in this issue covered everything from cable design, optical-fiber transmission, splicing equipment and the use of semiconductor materials that were more advanced than silicon. One of the articles featured the application of fiber optics in offshore systems. This type of environment posed tougher requirements on fiber design in terms of safety, weight and transmission distances.

The age of the internetIn 1998, an entire issue of

Ericsson Review was dedicated to the internet and its impact on different aspects of the telecoms industry. Scalability was the name of the game as were solutions that optimized the use of network resources. The focus of the articles in this issue was access, as at the time, most of the costs were hidden in this part of the network. In the wake of liberalization, the access network was deemed to be the battleground for competing offerings in terms of bandwidth and service levels.

A new millennium In 2007, Ericsson Review

carried an article about LTE-SAE (long term evolution-system architecture evolution). The simplified and optimized architecture uses a minimum number of nodes in the user plane. In addition, new features have been introduced to simplify operation and maintenance.

Ericsson Review, F issue, 1984.

Ericsson Review, special internet issue, 1998.


Carrier Wi-Fi: the next generationBy controlling whether or not a device should switch to and from Wi-Fi, and when it should switch, cellular operator networks will be able to provide a harmonized mobile broadband experience and optimize resource utilization in heterogeneous networks.

certain conditions, it is particularly appropriate for handling spikes in data traffic. But to work well, it needs to be integrated.

So, the main factors that next gener-ation Wi-Fi will be able to capitalize on are:

the vast amount of unlicensed spectrum that can be used by this technology without the need for any regulatory approval;its ability to offload data traffic, complementing existing indoor solutions, such as small cell and DAS;the near-ubiquitous device support for this technology – including UEs that are non-cellular;the evolution of small cells to support both cellular and Wi-Fi technology; anda new level of maturity – exemplified by the development of, and device support for, new standards and products, such as Hotspot 2.0, 802.11ac, EAP authentication and additional solutions currently being defined, such as S2a mobility, and 3GPP/Wi-Fi integration.

the available resources. In a heteroge-neous network, resources can include 2G, 3G, LTE, macro, small cells, different carriers, different protocols (802.11g, 802.11c) as well as different channels. In addition, devices do not take into account the activity of other UEs, and so each decision to switch to or from a Wi-Fi network is made independently, and without any consideration for load balancing. In short, current practice is inherently inefficient.

BackgroundThe smartphone revolution and the near-ubiquitous support for Wi-Fi of modern devices have created new busi-ness opportunities and new challenges for telcos. Operators have so far deployed over 2 million Access Points (APs) in pub-lic spaces, and there are currently about 8 million hotspots worldwide. But why is there a need to integrate Wi-Fi into operator networks? The simple answer is that this technology is a good com-plement to existing solutions, and in

RU T H GU E R R A


AES Advanced Encryption StandardAKA Authentication and Key AgreementANDSF Access Network Discovery and Selection FunctionAP Access PointBSC base station controllerBSS basic service setCSMA carrier sense multiple accessDAS Distributed Antenna SystemDPI deep packet inspectionEAP Extensible Authentication ProtocolGTP GPRS Tunneling Protocol

HLR home location registerHSS Home Subscriber ServerLI Lawful InterceptionMME Mobility Management EntityMS-CHAP Microsoft’s Challenge- Handshake Authentication ProtocolMWC Mobile World Congress NAS non-access stratumPLMN Public Land Mobile NetworkRAT radio-access technologyRF radio frequencyRNC radio network controller

RRM Radio Resource ManagementRSRP reference signal received powerRSSI received signal strength indicator SON self-organizing networksTCP Transmission Control ProtocolTLS Transport Layer SecurityTTLS Tunneled Transport Layer SecurityUE user equipmentUSIM Universal Subscriber Identity ModuleWIC Wi-Fi controller WLAN wireless local area network

Next generation carrier Wi-Fi will overcome existing coordination issues in multi-RAT environments to become an integrated component of mobile broadband offerings. Guaranteeing the best mobile broadband experience and ensuring that resources in a heterogeneous network that includes Wi-Fi are utilized in an optimal way, is only possible if subscribers are connected to Wi-Fi when this is the best option for them and for the entire network. While this may sound obvious, the way subscribers currently switch to and from Wi-Fi is not optimal. Today, the decision to connect to Wi-Fi is taken by the device according to one basic principle: if Wi-Fi is available, then use it for data traffic.

However, this approach is short-sight-ed because it does not take into consid-eration real-time information about all

48


Switching the smart way

The results of a survey1 carried out with 24 service providers are illustrat-ed in  Figure 1. These highlight the challenges operators are focusing their attention on when it comes to carrier Wi-Fi deployment.

Some steps have already been taken to include Wi-Fi in mobile broadband solutions, such as EAP authentication. Some solutions are already supported by UEs, while others will be available short-ly. But much more can be done.

With these challenges in mind, the top three priorities for next generation carrier Wi-Fi are:

traffic steering 3GPP/Wi-Fi – to maintain optimal selection of an access network so quality of experience can be ensured and data throughput maintained; authentication – to provide radio-access network security for both SIM- and non-SIM-based devices; andDPI, support for unified billing and support for seamless handover – achieved by integrating with the core infrastructure already deployed for 3GPP access.

When the varied set of resources in a heterogeneous network can be com-bined and optimized, networks can provide increased capacity and the per-formance needed to give subscribers the desired level of user experience. So for Wi-Fi, the objective is not to turn it into a 3GPP technology, but rather to figure out how to add 3GPP intelligence and control over Wi-Fi usage, so that all resources are used in an optimal way while delivering the best user experi-ence. When Wi-Fi becomes just another RAT, the synergies and application of mobile network capabilities, intelli-gence and infrastructure will remove the burden from Wi-Fi to meet all of the challenges outlined in Figure 1.

With the operator in control, and with Wi-Fi networks that are integrat-ed with mobile radio-access and core networks, subscribers will experience high- performing mobile broadband that operates in a harmonized way. Operators will be able to control, pre-dict and monitor the choice of connec-tivity, allowing them to optimize both the user experience and resource utili-zation across the entire network.

Traffic steering 3GPP/Wi-FiIn a heterogeneous network, the type and amount of resources that are avail-able to provide mobile broadband are quite diverse, as networks are built using:

multiple technologies including GSM, WCDMA, LTE and Wi-Fi;several types of cells including macro, small cells and APs; andvarying network capabilities including carrier aggregation; different carrier bandwidths; 802.11n; and 802.11ac.

To provide the best user experience across all available resources and opti-mize resource utilization, the decision of whether or not to switch to Wi-Fi or back to cellular, and when to switch, needs to be made according to a more complex set of principles than simply: if there is an available Wi-Fi with ade-quate signal strength, then switch to it. The decision should take into account all available technologies and carriers, all visible cells, network and UE capa-bilities, radio conditions for each spe-cific UE (which requires decisions to be made in real time, as these conditions

fluctuate rapidly), and should ultimately be based on a calculated performance estimate for each UE, given the aggrega-tion of all terminals in the area.

In a heterogeneous network it is more desirable to move users to different tech-nologies, carriers and layers. And so, real-time coordination of all resources adds an additional layer of complexity to the Wi-Fi/cellular decision.

For example, if we assume that the overall goal is to obtain the best UE throughput (as this has a direct impact on user experience), UEs that have 3GPP carrier aggregation capabilities should switch to Wi-Fi much later than termi-nals that don’t.

In another scenario involving three UEs, the best solution is to assign the first UE to a coordinated macro/small cell to deliver higher throughput, to assign another 3GPP technology or car-rier to the second UE, and to allocate a Wi-Fi access point to the third device – as this UE has 802.11ac capabilities and a good radio link budget in Wi-Fi. Good coordination is required to make such a decision, so that the best solution is attained for all three UEs given

FIGURE 1 Mobile and Wi-Fi network integration: the main challenges

n= 24Challenges

46%

42%

38%

0 20 40 60 80 100

50%

50%

46%

58%

63%

67%

67%

79%

83%83%Supporting seamless handoverbetween mobile and Wi-Fi networks

Mobile data offload traffic steeringbetween mobile and Wi-Fi networks

Supporting QoS

Maintaining data throughput

Radio access security

Supporting seamless handoverbetween Wi-Fi access points

Supporting SIM and non-SIM devices

Integrating carrier Wi-Fi with3G/4G small cells

Authentication security

Deep packet inspection and policy control

Supporting unified billing acrossmobile and Wi-Fi networks

Generating revenue from Wi-Fi investment

49


SON features, information about poten-tial mobility decisions, as well as an awareness of all UEs in the area.

For years, operators have deployed mobility features both within and between 3GPP radio access networks that allow the network to determine in what cell, with which carrier or by what technology a UE should be hosted. These features make good use of all available resources to provide every-one connected to the network with an optimal mobile broadband experience. Decisions are made by the network on the basis of the available resources, radio conditions of UEs and the current state of the network (an aggregation of information from the network and all the UEs in the area).

A mobility featureEricsson’s concept of a mobility fea-ture for switching to and from Wi-Fi networks provides operators with the capability to control when UEs connect to Wi-Fi and when data traffic should be switched back to 3GPP. This concept is based on the forecast performance for the specific UE.

By deploying a mobility feature that is based on performance, operators will be in a position to add Wi-Fi to their mobile

broadband resources, just like any another RAT. And as the same entities make the decisions about 3GPP mobili-ty, 3GPP/Wi-Fi mobility and distributed SON features, decisions are coordinat-ed, preventing devices from looping, and improving resource utilization as a result of less signaling.

To ensure that the benefits of a mobil-ity feature supporting real-time traffic steering can reach mass market quick-ly, such a feature needs to be developed to work on legacy UEs without the need to install a client. To achieve this and to avoid placing additional requirements on UEs that would delay the applicabil-ity of the feature, an interface needs to be placed between the controllers – BSC, RNC, Wi-Fi controller (WIC) and MME – of the different technologies. Switching decisions need to be made jointly by these controllers, as illustrat-ed in Figure 2 – a simplified flow dia-gram for a performance-based mobility feature.

To illustrate how the concept works, it may be useful to consider the case of a UE that is moving closer to an operator AP, but is currently hosted in a 3G cellu-lar network:

the default UE behavior is to request a connection to the Wi-Fi on detection of the AP;the WIC will proxy the authentication request; on authentication, the WIC provides the RNC currently hosting the UE with an estimated performance for the UE in Wi-Fi; if the 3G throughput estimation is higher than the estimates for Wi-Fi, the RNC will order the WIC to reject the connection, thus maintaining data transmission over 3G.

As long as the UE is in the vicinity of the AP, the connection manager in the UE will keep trying to connect to the Wi-Fi, allowing the network to decide to switch when and if the conditions become more favorable in Wi-Fi.

Once a client is accepted into the Wi-Fi access network, the UE throughput in Wi-Fi and the estimated throughput for the UE in the NodeB are monitored con-tinuously. When NodeB throughput sur-passes Wi-Fi performance for the UE, it is disconnected from Wi-Fi and data communication will switch back to 3G.

their individual requirements and those of the other two devices.

The ability to assure predictable and consistent switching behavior in het-erogeneous network architectures eas-es network planning, makes optimal use of available resources and ensures good performance for all subscribers independent of device. This implies that a Wi-Fi/cellular decision-making mechanism needs to be independent of UE type, UE operating system and UE vendor.

Crystal ballAs good predictions can prevent UEs from switching too often from one RAT to another, the ability to forecast net-work states and available capacity is a fundamental tool for the RAN to oper-ate effectively and to help improve the mobile broadband experience. Avoiding the ping-pong effect is beneficial because the less switching among RATs, the fewer the signals between the RAN and the core, and the less the impact on battery consumption in the terminal.

The RAN is the best place in the net-work to measure available resources for specific UE forecasts, as this part of the network has a complete overview of resources, knowledge of the distributed

FIGURE 2 Concept of a performance-based mobility feature

eNodeB

AAA

MMEMME

BSC

RNC

WIC

Request to connect Authentication

Accept/reject

UE throughput estimate for Wi-Fi


Accept/reject after comparison with UE estimated throughput in 3G Accept/reject after

comparison with UE estimatedthroughput in 2G

Accept/reject after comparison with UE estimated throughput in LTE


50



Adding smartnessIn addition to guaranteeing the best experience for users and optimizing use of resources, maintaining control over the Wi-Fi/cellular switching process pro-vides operators with a platform that can be rapidly adapted to include additional parameters. For example, operators will be able to include subscription parame-ters or service considerations into the switching decision, without having to wait for support to be implemented by all UEs.

The real-time switching decision has a greater impact on user experience for UEs that are in connected mode. Consequently, an extension of existing mobility features for connected mode that includes support for Wi-Fi would optimize and enhance this decision. This could be achieved by including Wi-Fi parameters in UE measurement reports (which currently include only cellular measurements) and so further standardization and extension of mobil-ity mechanisms is foreseen.

The way these measurements would be handled by the UE is illustrated in Figure 3, and the suggested traffic-steering process would work as follows:1. the RAN instructs the UE by setting

thresholds or conditions on which UE should perform measurement reporting. For example, the RAN may provide a condition based on signal strength – RSSI – broadcasted WAN metric and load or received power – RSRP;

2. if any of the thresholds are met, the UE reports WLAN measurements back to the RAN;

3. on the basis of this information, the current network state and any other information it has available, the 3GPP RAN will send a traffic-steering message to the UE to switch data traffic to or from WLAN.

The 3GPP mobility mechanisms for idle mode will be extended in 3GPP Rel-12, providing coordination between 3GPP and Wi-Fi for UEs in idle mode.

While the suggested mobility feature can evolve to include enhancements, it can be included in any vendor solution without requiring any direct interfaces, thus increasing the level of coordination between the various access networks, cell types and carriers from different vendors.

Traffic steering today – limitations Today, the decision to move traffic between 3GPP and Wi-Fi is made independently by each UE based on the UE’s implementation of the connection manager. With varying implementations, UEs from differ-ent vendors are likely to carry out the switching decision at different times. This makes the ecosystem unpredict-able, increases capex and opex and limits the operator’s ability to guar-antee superior user experience.

The algorithm currently used by UEs is oversimplified: the selection of Wi-Fi over cellular is based sole-ly on the availability of Wi-Fi. Once a UE detects Wi-Fi, it automatically shifts data traffic to it. Even if devic-es become smarter, UEs will still make decisions based on a single cell and a single AP – decisions that are not coordinated with the network (which can see all resources) or with other UEs in the area. Most of the UEs in the coverage area of an AP will therefore try to connect to it.

What is Wi-Fi good at?Let’s consider how Wi-Fi is designed: to provide high peak rates and low latency for a limited number of users in the vicinity of an AP. When too many users are connected, or when the users have bad link bud-gets, the performance of the carrier

sense multiple access (CSMA) protocol degrades and the total capacity of the AP drops rapidly.

To avoid this type of degradation, an AP should not permit a UE to attach unless the following conditions are met:

the service level provided to the UE over Wi-Fi is better than the level offered over cellular; andthe quality of the Wi-Fi connection is good enough to ensure that the experience level for users that are already connected to the AP will not be overly affected by the addition of another user.

Proof of conceptBy putting the network in control, and using all the network information and aggregated information from UEs in the decision-making process, the net-work can provide the best user expe-rience as well as increasing its overall mobile broadband capacity. The proof of concept measurements for such an approach are shown in Figure 4.

The left-hand graph shows that by rejecting users with bad link budgets and poor radio conditions, the APs can deliver higher throughput rates with exactly the same equipment. Without the mobility feature, the AP reaches a maximum peak rate of 30Mbps, and delivered throughput for most of the UEs is between 0 and 17.5Mbps. With the feature activated, the same AP reaches a maximum peak rate of

FIGURE 3 Connected mode mobility to and from Wi-Fi

3GPP RAN3GPP RAN

33

22

11

WLANWLAN

BSS/WANBSS/WAN

51


UEs and the decisions made can lead to devices ping-ponging from one access technology to another, causing signal-ing storms and additional battery con-sumption – leading to a degradation in overall network performance.

Many implementationsThe variety of connection manager implementations in UEs and the range of radio sensitivities means that the out-come of today’s oversimplified mobili-ty mechanism is unpredictable, which makes it difficult for operators to plan and optimize networks.

Approaches working against each other Today, 3GPP mobility features are deployed both within a radio access technology and among technologies. These features are controlled by the RAN, and work in real time to provide the desired mobile broadband expe-rience and optimize resource man-agement. They also take into account network parameter changes, as SON features are triggered to adapt network behavior to constantly changing sub-scriber activity.

If the 3GPP/Wi-Fi decision continues to be made by the UE, this can interfere

40Mbps, with most users receiving between 20 and 40Mbps, and the capac-ity is increased up to 100 percent.

The right-hand graph shows that without the mobility feature, UE per-formance drops by up to 20Mbps; 50 percent of users experience a drop of 10Mbps and all of them experience inferior performance when moving to Wi-Fi. With the feature activated, the throughput difference in the worse case is 7.5Mbps and on average the difference is zero – thus mobile broadband experi-ence is maintained.

Network probingAnother approach to optimizing con-nections is to force the UE to probe the network before it switches. However, this approach increases the uncertain-ty of measurements – the greater the number of UEs probing the network, the more unrealistic measurements become. Probing can at best provide an indication of the level of performance available at a given moment, and results cannot take into consideration changes to network parameters created by SON features, triggering of mobility features or resource utilization by other UEs. The lack of coordination among individual

with deployed 3GPP mobility mecha-nisms, causing devices to loop among access networks and wasting resourc-es. For example, an active 3GPP mobil-ity feature might be in the process of switching a UE to another technology (from LTE to 3G, for example) when the device decides to switch to Wi-Fi instead. If the Wi-Fi becomes overloaded, it will switch the UE back to the 3GPP net-work, which might trigger an addition-al switch to another cell.

Integration toolsSome complementary tools, such as Hotspot 2.0 and ANDSF, have been devel-oped to ease the integration of Wi-Fi.

By using HS2.0, broadcasted BSS load and WAN metrics, UE-based switching improves somewhat2, as these mecha-nisms can prevent UEs from connect-ing to overloaded or limited backhaul Wi-Fi APs.

However, a low load does not neces-sarily imply that the resource availabil-ity of an AP is good. Low load can be the consequence of a bad radio environment caused by interference. To use these mechanisms efficiently, the refresh frequency of broadcasted load update needs to be fine-tuned to prevent mass

FIGURE 4 Measured capacity increase by RRM and coordination of Wi-Fi and 3GPP

0000

1010

2020

3030

4040

5050

6060

6060

8080

9090

100100

2.52.5 55 7.57.5 1010 12.512.5 1515 17.517.5 2020 22.522.5 2525 27.527.5 3030 32.532.5 3535 37.537.5 4040 42.542.5

Throughput in Wi-Fi(Mbps)

Throughput in Wi-Fi(Mbps)

C. D. F.(%)C. D. F.(%) Legacy AP

Network controlled APLegacy APNetwork controlled AP

-20.0-20.000

1010

2020

3030

4040

5050

6060

6060

8080

9090

100100

-17.5-17.5 -15.0-15.0 -12.5-12.5 -10.0-10.0 -7.5-7.5 -5.0-5.0 -2.5-2.5 00 2.52.5 5.05.0 7.57.5 1010

Throughput difference after selecting Wi-Fi(Mbps)

Throughput difference after selecting Wi-Fi(Mbps)

C. D. F.(%)C. D. F.(%) Legacy AP

Network controlled APLegacy APNetwork controlled AP

19.33.2019.33.20 19.35.0019.35.00 19.36.4019.36.40 19.38.2019.38.20 19.48.2019.48.2019.40.0019.40.00 19.41.4019.41.40 19.51.4019.51.4019.46.4019.46.40 19.56.4019.56.4019.43.2019.43.20 19.53.2019.53.2019.45.0019.45.00 19.55.0019.55.0019.50.0019.50.00

00

2525

5050

TimeTime

Throughput(%)Throughput(%)

52



toggling between cellular and Wi-Fi accesses as a result of brief load peaks.

The ANDSF mechanism augments PLMN selection, providing operators with improved control over the deci-sions made by devices. Through a set of static operator-defined rules, ANDSF guides devices through the decision-making process of where, when and how to choose a non-3GPP network. But the final decision is still made by the UE and is still dependent on the imple-mentation of the device’s connection manager. ANDSF is defined to take into account the local environment – in oth-er words, the specific connection man-ager implementation in the UE – and so UEs from different vendors behave in different ways. Using ANDSF might lead to an unpredictable outcome.

Even if they are enhanced with radio information, NAS-based solutions do not have the real-time information needed to make the optimal decision. This is because requirements on infor-mation updates would become too demanding for terminals and networks.

The ANDSF mechanism is, however, valuable as it provides operators with some level of control over the 3GPP/Wi-Fi selection process, and it takes into account static or slowly changing network parameters. However, this mechanism has an impact on battery utilization in UEs, as it scans continu-ously for Wi-Fi and does not cater for the rapidly fluctuating RF environment experienced by a mobile device. Neither does the mechanism take into account UE and radio access capabilities (such as LTE, HSPA, EDGE, carrier aggregation, cell bandwidth, 802.11n and 802.11ac), or network changes (as a result of self-optimizing features in RAN). And it does not provide real-time coordination with decisions in other UEs in the area.

By deploying an HS2.0- and/or ANDSF-only approach, the same policy will be applied by all UEs configured with that policy, independently of their RF condi-tions, cell and network capabilities in a manner that is not coordinated with other UE decisions or network features. This results in an inability to guarantee the best user experience as well as cre-ating additional opex and capex for the operator, as resource utilization is not optimized on a network level.

Scenario: many users on a trainConsider this scenario: a train is entering a station where two APs are deployed. There are 200 subscribers on the train with active mobile broadband sessions.

The APs in the station perform best when the number of UEs attached is less than 20. If AP1 has a slightly better sig-nal strength than AP2, it is possible that 180 of the UEs on the train will try to connect to AP1, while the remaining 20 UEs will select AP2. This is because each UE is ignorant of the decisions made by all the other UEs. Subsequently, a number of UEs connected to AP1 will be steered to AP2 or back to 3GPP as a result of the high load on AP1.

When two UEs apply the same pol-icy, the one that supports carrier aggregation should move data traffic to Wi-Fi much later – which requires coordination.

So while ANDSF can be used as a sim-ple mechanism for offloading, it neither guarantees the best user experience nor optimal resource utilization.

Seamless authentication and increased securityTo secure the air link between a UE and a hotspot, Passpoint devices use the WPA2 Enterprise security protocol. This is a four-way handshaking protocol that is based on AES encryption, and offers a level of security that is comparable to cellular networks.

The Hotspot 2.0 specification sup-ports four commonly deployed stan-dard protocols:

SIM-based authentication – EAP-SIM for devices with SIM credentials and EAP-AKA for devices with USIM credentials; andnon-SIM-based authentication – EAP-TLS for client and server-side authentication, with a trusted root certificate, and EAP-TTLS with MS-CHAPv2 for user-name-password authentication.

The Hotspot 2.0 specification comple-ments WPA2 Enterprise security by incorporating features, such as layer-2 traffic inspection and filtering, as well as broadcast/multicast control, which are often used to mitigate common attacks on public Wi-Fi deployments.

Seamless authentication (where users are not required to introduce a user name and password) and increased security are key to Wi-Fi usage. This phenomenon is illustrated in Figure 5, which shows the uptake of Wi-Fi in an airport network enabled with EAP-SIM/AKA authentication.

By reusing SIM mechanisms, sub-scribers only need to be provisioned once in the HSS/HLR to use an opera-tor’s mobile broadband.

Core integration and IP session mobility in 3GPP/Wi-Fi Connection to the mobile core net-works from Wi-Fi equipment

FIGURE 5 Usage of Wi-Fi shown by authentication method

Jan2011

Mar2011

May2011

Jul2011

Sep2011

Nov2011

Jan2012

SIM sessionsWeb sessions

Mar2012

May2012

Jul2012

53


broadband by overcoming the limi-tations of two IP stacks (one for cellu-lar and one for Wi-Fi ) used in current UE implementation. The double stack approach results in IP address realloca-tion every time devices move between cellular and Wi-Fi.

Some applications have tried to overcome these limitations by deploy-ing tunneling mechanisms and using buffers, so that users do not notice the delay introduced in the session by the IP address reallocation (reestablishment of TCP sessions, socket reallocation, packet loss and rerouting delays). But these solutions are not optimal for real-time applications such as video stream-ing or voice calls; nor do they support IP-address-dependent security mecha-nisms such as VPN services, and they are not supported by enough UEs to be mass-market. By moving the support for IP address continuity to the modem, it is possible for a device to keep its IP address at handover in a way that is com-pletely transparent to applications. This is what 3GPP Rel-12 offers.

ConclusionNext generation carrier Wi-Fi address-es the technical challenges relating to mobile broadband Wi-Fi. By enabling operators to add Wi-Fi capacity where 3GPP spectrum is scarce, Wi-Fi-based business models can be included into operator offerings to maintain optimal mobile broadband experience.

With operators in control and Wi-Fi integrated into heterogeneous net-works, the mobile broadband experi-ence will become harmonized and provide users with the best possible performance. By taking control, opera-tors will be able to predict and monitor the choice of connectivity, maintaining the user experience and optimizing net-work resource utilization.

Ericsson will develop its 3GPP and Wi-Fi portfolios based on the concepts outlined above.

gives subscribers direct access to oper-ator services, and provides operators with a mechanism to improve visibili-ty, reuse their mobile broadband infra-structure, and integrate with different systems (LI, DPI, session establishment and management, policy decision and enforcement, reuse of wholesale and roaming agreements, access to operator branded or hosted services, and online and offline charging). Such an architec-ture provides a harmonized platform for handling cellular and Wi-Fi access, allowing Wi-Fi to become a continua-tion of the operator’s mobile broadband experience.

Packet-core network integration helps operators to gain control of non-cellular traf fic, and consequently of the user experience.

Seamless handover – uninterrupted IP session mobility between Wi-Fi and 3GPP – is a key technology when it comes to providing good user expe-rience and mobility. To implement it, Ericsson has taken a leading role in both the standardization process and in prod-uct development.

In 3GPP Rel-11, the GPRS Tunneling Protocol (GTP) – which is widely deployed in mobile networks – was spec-ified for the S2a interface. Use of this interface, for trusted non-3GPP access, makes it possible for UEs to connect to the Wi-Fi network and utilize packet-core network services without mobility or tunneling support in the UE. The S2b interface was also included in the stan-dard for connection to the core through untrusted Wi-Fi.

3GPP Rel-12 standards will enable GTPv2-based IP session mobility between Wi-Fi and 3GPP over the S2a interface. Ericsson together with Qualcomm demonstrated how this works at MWC 2013, and IP session mobility is expected to appear in ven-dor products in 2014.

The Rel-12 solution will bring the user experience even closer to mobile

1. Infonetics Research, May 2013, CarrierWiFiOffloadandHotspotStrategiesandVendorLeadership:Global Service Provider Survey, available at: http://www.infonetics.com/pr/2013/WiFi-Offload-and-Hotspot-Strategies-Survey-Highlights.asp

2. Ericsson, December 2012, Ericsson Review, Achieving carrier-grade Wi-Fi in the 3GPP world, available at: http://www.ericsson.com/news/121205-er-seamless-wi-fi-roaming_244159017_c?idx=10&categoryFilter=ericsson_review_1270673222_c

References

Ruth Guerra

joined Ericsson in 1999 and is currently strategic product manager for Wi-Fi integration at Product Line

Wi-Fi and Mobile Enterprise. She is an expert in Wi-Fi and 3GPP networks, focusing on the strategic evolution of Wi-Fi and 3GPP RAN integration. She holds an M.Sc. in Telecommunications from the Technical University of Madrid (Universidad Politécnica de Madrid), Spain.

54



Ericsson SE-164 83 Stockholm, SwedenPhone: + 46 10 719 0000

ISSN 0014-0171 284 23-3220 | Uen

Edita Bobergs, Stockholm© Ericsson AB 2014