EDGE for the masses Uncertainty drives SDR ACM …iopp.fileburst.com/we/we_01_47.pdfEDGE for the...

I S S U E 4 7

T E C H N I C A L I N S I G H T F O R W I R E L E S S P R O F E S S I O N A L Sw i r e l e s s . i o p . o r g

O C T O B E R / N O V E M B E R 2 0 0 6

EDGE for the masses Uncertainty drives SDR ACM boosts backhaul

Open specifications ease installation

WEOctCover_1 5/10/06 12:04 Page 1

http://wireless.iop.org

CST – COMPUTER SIMULATION TECHNOLOGY | www.cst.com | [email protected]

CHANGING THE STANDARDS

So dominant. So flexible.The evolution in 3D EM simulation software. CST MICROWAVE STUDIO® 2006.

� Achieve the impossible. Use a toolthat dominates its territory andoffers the flexibility to choose theright solver at the right time, withinone interface.

Discover the latest evolution in 3D EM simulation software andexperience CST MICROWAVE STUDIO®[CST MWS]’s complete technology.

CST MWS is the first commercial HF 3D EM code to offer the advantagesof time and frequency domain,hexahedral and tetrahedral meshing,united in one interface.

Choose the technology best suitedto your structure. Embedded in an advanced design environment,CST MWS can be coupled with all

CST STUDIO SUITE™ solver techno-logy including circuit and thermalsimulation.

CST MICROWAVE STUDIO® is thepowerful market leading time domaintool for 3D EM simulation. It has bene-fited from over 20 years experiencein the area of numerical field calcu-lation,and is used by industry marketleaders worldwide.� Change your outlook on simulationtechnology. Choose the flexibility of complete technology with CSTMICROWAVE STUDIO®.

Document2 24/5/06 15:59 Page 1

http://www.cst.com

mailto:[email protected]

L E A D E R 5IMS to the rescue

N E W S 7Femtocells could stall dual-mode voice-over-WiFi

A N A LY S I S 11Product Development: Mobile success comes aftermany hurdlesApplications: Mobile application testing is notoptional

O P I N I O N 15Graham BellThe way that handsets are marketed to the public isstifling innovation.

E D G E 16Dual-transmitter architecture brings EDGE to the massesThe integration of GSM and EDGE transceivers isenabling high-volume handset production.

O P T I M I Z A T I O N 18Multiple technologies bring optimization challengesLiberalization of the telecoms industry is bringingnew technologies onto existing base-station sites,further complicating network-optimization.

I M S 20Overcoming challenges on the road to IMSHow should operators address the challenges ofachieving smooth migration to IMS?

B A S E S T A T I O N S 22A new era of base-station developmentOpen specifications for the interfaces betweenmodules is easing base-station development.

P R O D U C T F O C U S 25ACM is the key to next-generation backhaul ●Mobile IP services demand greater access speeds

P R O D U C T S 30Test and measurment

T H E F U T U R E 31Uncertainty is driving SDRSoftware-defined radio is poised to change cellulartechnolgy, explains Steve Jennis, senior vice presidentfor corporate development at PrismTech.

3C O N T E N T S

OCTOBER/NOVEMBER 2006ISSUE 47

Institute of Physics Publishing Ltd,Dirac House, Temple Back, BristolBS1 6BE, UK.Tel: +44 (0)117 929 7481Editorial fax: +44 (0)117 925 1942Advertising fax: +44 (0)117 930 1178Web: wireless.iop.orgE-mail: [email protected]

EDITORIALEditor: Hamish Johnston;[email protected]: Paul JohnsonArt director: Andrew GiaquintoTechnical illustrator: Alison Tovey

ADVERTISINGSales manager: Simon Allardice;[email protected] co-ordinator: Jayne Boulton;[email protected]

MARKETING AND CIRCULATIONProduct manager: Angela Peck;[email protected]

PUBLISHERSarah Chilcott; [email protected]

ISSN 1471-3888

SUBSCRIPTIONSFree to qualifying readers (see subscription cardenclosed or register online at the above Webaddress). Subscriptions £96/7139 ($147 in theUS and Canada) for six issues to readers who donot meet qualifying criteria. Orders to WirelessEurope, IOPP Magazines, WDIS Ltd, Units 12 &13, Cranleigh Gardens Industrial Estate, Southall,Middlesex UB1 2DB, UK. Tel: +44 (0)20 86067518; fax: +44 (0)20 8606 7303; e-mail:[email protected]

©2006 IOP Publishing Ltd

The contents of Wireless Europe do not representthe views or policies of the Institute of Physics, itscouncil or officers unless so identified.

Printed by Warners (Midlands) plc, Bourne,Lincolnshire, UK.

BPA Worldwide Business Publication Audit

membership applied for January 2005

w i r e l e s s e u r o p e wire less. iop.org O c t o b e r / N o v e m b e r 2 0 0 6

Achieving a smooth transition to IMS. p20.

Bell: “Handset marketing stifles innovation.” p15.

Multiple technologies push antenna limits. p18.

On the cover:Open specificationsare easing base-stationdevelopment. p22.

WEOctContents_3 5/10/06 16:47 Page 3









In Europe, call 44-2392-232392 > ISO 9001 certified Visa/MasterCard accepted (except in Europe)800-411-6596 > www.anaren.com

Components engineered for what will be.

Was.

Is.

Xinger®0805, low-profilepower dividersAt last, better performance than printed andlumped-element solutions – in a low-profile, high-volume SM package! Our subminiature-sizedWilkinson power dividers offer: > Low-profile, sub-miniature SM format,

only .079 x .049 x .021” (2 x 1.25 x .53mm)> Models covering 400-8000MHz> Models suitable for GSM, WCDMA, 802.11 b, g;

MIMO b, g; Bluetooth®; Zigbee> Good output port isolation and return loss > Input/output impedance of 50Ω and 75Ω> Insertion loss as low as 0.6dB typ> Non-conductive surface> Proven, high-volume manufacturing capability

Broadcast! Models available for DVB-H (Europe & US) and DVB-S

RoHS compliantCall for free samples or our complete CCG engineering kit (qualified inquiries).

Xinger®0805, ultra-low profile hybrid couplers3dB couplers don’t get any smaller than this! Our new Hybrids are ideal for balanced power and low-noise amplifiers – and anywhere low-insertion loss and tight amplitude/phase balance are required:> Ultra-low profile, SM package measures a mere

only .079 x .049 x .028” (2 x 1.25 x .7mm)> Models covering 1700-3600MHz > Designed for PCS, DCS, DECT, WCDMA-3G jobs> Insertion loss as low as 0.3dB > Isolation as high as 22dB > Port impedance of 50Ω> Proven, high-volume manufacturing – shipped

on tape and reel

RoHS compliantCall for our free samples or our complete CCG engineering kit (qualified inquiries).

Xinger®0404, nano-profile balunsWhile matching ceramic baluns on performance,our Xinger-brand unbalanced to balanced transformers are a “must” for modern, consumerapplications. Featuring:> Nano-profile, sub-miniature SM format,

only .039 x .039 x .026” (1 x 1 x .65mm)> Models covering 2400-5900MHz> Compliance with 802.11 a, b, g; MIMO b, g;

Bluetooth®; Zigbee> Low losses: 0.6dB typ> Typical return loss: >16dB> Wide variety of differential impedances> Non-conductive surface> Proven, high-volume manufacturing capability

Modular! Choose from dual, triple, and quad configurations!

RoHS compliantCall for our free samples or our complete CCG engineering kit (qualified inquiries).

Wireless Europe 906.qxp 9/18/06 9:34 AM Page 1

http://www.anaren.com

5L E A D E R


N E W S I N B R I E F

HSUPA test is a firstAnritsu has released the first 3GPP-approved protocol conformance test casefor the High Speed Uplink Packet Access (HSUPA) enhancement to the UMTS3G air interface. Anritsu claims that “this represents a huge step forward,enabling the mobile industry to deliver commercially available HSUPAterminals”. HSUPA boosts data rates in the uplink from 384 kbit/s to5.7 Mbit/s. The TS34.123 test case 8.2.6.50 is part of the 3GPP Release 6specification. It was submitted by Anritsu to the 3GPP RAN-5 working groupfor approval in early September.

RFMD will sell Jazz interestsRF Micro Devices (RFMD) is selling its financial interest in Jazz Semiconductorfor approximately $25 m. Jazz is a US-based chipmaker that produces RF andmixed-signal devices for mobile phones. It specializes in silicon and silicongermanium fabrication processes. RFMD bought a $60 m minority stake inJazz in 2002 to secure a supply of low-cost silicon wafers for its single-chiptransceivers. It will continue joint development projects with Jazz and in 2007the foundry will begin to produce RFMD’s next-generation single-chip EDGEtransceiver. On completion of the sale, Jazz, which was spun out of ConexantSystems in 2002, will be wholly owned by the holding company Acquicor.

RFI moves on North AmericaRFI Global Services has opened an office in Dallas, Texas, to improve thedelivery of its 2G/3G cellular approval services to mobile-phone and chipsetmanufacturers in the US and Canada. The North American effort will beheaded by Kenwood Martin, who will use RFI’s certification and validationexpertise to help manufactures to ensure that their products can be soldworldwide. RFI is based in the UK.

3G users will top 1bn by 2010There will be more than 1 bn users of 3G mobiles phones by 2010, predicts areport from the analyst firm Strategy Analytics. This represents significantgrowth from the 167 m people that are expected to use 3G services by theend of this year. Entitled “Worldwide cellular user rorecasts 2006–2011”, thereport also predicts that 3G will account for half of all mobile-servicerevenues by 2010, even though there are expected to be twice as many 2Gusers by that time. David Kerr of Strategy Analytics commented: “3G still hasits detractors but it is starting to make a significant contribution to thewireless market-place. Even in emerging markets it will generate more thanone quarter of service revenues by the end of this decade.”

BT and ZTE partner on 3G TVBritish Telecommunications (BT) and the Chinese equipment maker ZTE aredeveloping a 3G handset that can receive mobile TV services based on thedigital audio broadcast (DAB) standard. The handset will be used by BT in itsBT Movio mobile TV and digital radio service, which has just launched in theUK on the Virgin Mobile network. The Virgin service is currently based on theLobster 700TV 2G handset.

Argogroup to boost Chinese handset qualityArgogroup and the Chinese Academy of Telecommunication Research (CATR)have formed a strategic alliance to boost the quality of handsets exportedfrom China. The five-year agreement will see UK-based Argogroup’s DeviceMaster handset-monitoring technology forming the basis of China’s newHandset Quality Export Promotion project. According to Xie Yi, chairman ofCATR’s Terminal Testing and Approval Forum, “[Argogroup] is the only vendorwho has really solved the handset profiling challenge”.

It’s been a good five or six years since 3G licences were issuedthroughout most of Western Europe and services have beenlive for at least two years in nearly all major markets, but isanyone actually using 3G services?

As of June, fewer than 10% of Vodafone’s Europeancustomers were using a 3G phone, with Italy leading thepack with 12% of its subscribers owning 3G handsets.Dedicated 3G operator 3 is doing better than Vodafone. Ithas about 3.8 m customers in the UK compared withVodafone’s 1.2 m.

And just because someone has a 3G handset doesn’t meanthat they’re using advanced data services. Vodafone launched3G services in February 2004 and the following June 1.8% ofits European revenues came from data services – notincluding SMS text messages. Two years later this has nearlytrebled to 4.9% – a significant increase, but surely not therevolution that the industry expected.

Operators like Vodafone have paid huge amounts ofmoney for 3G licences and networks. They’ve spent more ondeveloping services and subsidizing handsets. But theirsubscribers – me included – remain firmly in the late 1990sby only using voice and text services.

Of course, there’s always another technology round thecorner that could save the day for 3G operators – and the IPMultimedia Subsystem (IMS) is no exeception. IMSpromises to fix a key problem that has held 3G services back– the inability to deliver data services to a subscriberregardless of their access mode. This is why IMS is shapingup as a must-have for 3G service providers.

Avi Gabbay certainly thinks so and on page 20 he arguesthat IMS will have a profound impact on mobilecommunications because it will allow operators to developservices that can take full advantage of fixed and wirelessnetwork resources.

dotMobiWhile IMS is definitely a “big idea” that will take some timeto implement, the mobile world has made a small moverecently that could have a positive impact on mobile Internetuse. The .mobi mobile-Internet domain name was launchedin September by a joint-venture company called dotMobi(see page 7). While there’s nothing earth-shattering about anew domain name, dotMobi plans to support Webdevelopers with stylesheets, standards and tools that couldensure that mobile-Internet users enjoy a consistent andreliable service across all sites. This is something that doesn’treally happen today, which is why 3G hasn’t captured theimagination of European subscribers.

Hamish Johnston, Editor

IMS to the rescueE D I T O R I A LN E W S


Document4 13/7/06 12:39 Page 1

http://www.vectron.com

7U P F R O N T

More than 300 million dual-mode cellular/voice-over-WiFihandsets could be sold in 2011,but the rise of the femtocellcould have a negative impact ongrowth of voice-over-WiFi,claims ABI Research in its report“Voice Over Wi-Fi Handsets”.

“The 802.11n standard will bethe most commonly deployedtechnology in dual-mode hand-sets,” says analyst Philip Solis.However, he warned that thearrival of the femtocell base sta-tions in the next few years couldhave a significant negativeimpact on the success of voice-over-WiFi services.

Low-cost femtocell base sta-tions could provide cellular cov-

erage to offices and even homes.“As frequency reuse issues areresolved, femtocells will providesome counterbalance to the trendtowards dual-mode handsets,”said Solis.

This has already happened inthe UK, for example, where radiospectra has been made availablefor low-power indoor servicessuch as GSM.

Solis said that mobile operatorsare looking to deploy dual-modeservices based on the UniversalMobile Access (UMA), whileother telecoms providers are keenon using the more advancedSession Initiation Protocol (SIP).“It’s all about who controls thecustomer,” explained Solis.

“Mobile operators can use UMAto keep the customers who wantto use cheaper wireline minutes.It’s good for the end-user and it’sgood for the [mobile] operator,because it uses the customer’sown broadband connection tobackhaul traffic to the core net-work.”

UK-based ip.access hasrecently unveiled the femto3Gbase station, which is specificallydesigned for deployment in thehome and small businesses. Thefemto3G is also backhauled viathe customer’s broadband con-nection – through a conventionalADSL router. But, unlike UMAservices, it does not rely on dual-mode cellular/WiFi handsets.


Femtocells could stall dual-mode voice-over-WiFi

Freescale in privateequity buyout

Freescale Semiconductor is to bebought by a consortium of pri-vate firms in a deal worth$17.6 m. Led by the BlackstoneGroup, the consortium includesthe Carlyle Group, PermiraFunds and the Texas PacificGroup. The deal received unani-mous approval by Freescale’sboard of directors.

US-based Freescale was origi-nally the chip manufacturingarm of Motorola and was spunout in 2004. It has annual sales ofabout $6 bn and focuses on mak-ing components for telecommu-nications, automotive and otherapplications.

dotMobi hits theairwavesThe .mobi mobile-Internetdomain name was launched inSeptember to encourage mobile-phone users to access Internetsites. The launch is being co-ordinated by mTLD Top LevelDomain Ltd (also known asdotMobi), a joint-venture com-pany based in Dublin, Ireland,with offices in Washington, DC.

The .mobi sites are designedspecifically to be accessed bymobile phones and other portabledevices. To support the develop-ment of these sites, dotMobi hasestablished a set of best practices,style sheets, standards and tools.“In the past, using the Internet ona mobile device presented hugeobstacles for the consumer,” saiddotMobi chief executive NeilEdwards. “We’re creating a linkbetween mobile operators,Internet content providers andthe millions of mobile users seek-ing a consistent, reliable andenjoyable Internet experience.

dotMobi is backed by severalleading players in the mobiletelecoms industry includingNokia, Ericsson, Samsung,Vodafone and T-Mobile.

BenQ is closing its Germany-based mobile-phone businessand shifting its handset activitiesto Asia. Called BenQ Mobile, themobile-phone manufacturingbusiness was created in 2005when Taiwan-based BenQbought Siemens’ mobile-phonebusiness.

BenQ is terminating financialsupport of BenQ Mobile and thecompany is expected to file forinsolvency in a Munich courtwithin the next few days. BenQis cutting off its German sub-sidiary because the handsetmaker has not brought in suffi-

cient revenue. BenQ will con-tinue to offer products under theBenQ-Siemens brand, which waslaunched earlier this year.

These developments will affectabout 3000 BenQ Mobileemployees currently based inMunich and Bocholt. BenQMobile also has manufacturingoperations in China, Poland,Brazil, Mexico and Taiwan. Thefate of these sites is still underconsideration, according toBenQ. Earlier this year, BenQMobile sold its research anddevelopment centre in Aalborg,Denmark, to Motorola.

BenQ cuts off German subsidiary

The femto3G base station. Femtocell basestations deployed in homes andbusinesses could put the brakes on growthin dual-mode cellular/WiFi handsets.

Freescale’s Chairman of the Board MichelMayer approved the $17.6 million sale.

BenQ Mobile chief executive Clemens Joos has taken immediate action in order tosecure ongoing business operations at the German handset maker.

WEOctNews_7,8,9 5/10/06 06:40 Page 7


8 B U S I N E S S

O c t o b e r / N o v e m b e r 2 0 0 6 wire less. iop.org w i r e l e s s e u r o p e

Alcatel plans to buy Nortel’sUMTS radio-access infrastruc-ture business for $320 m. Nortelwill continue to produce GSMand CDMA radio-access equip-ment as well as to develop 4Gcellular technologies. It will alsocontinue to sell equipment forUMTS core networks. This is thesecond major North Americanacquisition by Alcatel of Francethis year, which is in the processof buying US-based LucentTechnologies.

According to Nortel’s chief

executive Mike Zafirovski, thecompany is selling because it hasfallen behind its competitors inthe UMTS radio-access market.“Our UMTS access businesslacks the scale and momentumneeded to become profitable,”said Zafirovski.

Radio-access technology is nolonger a priority at Canada-basedNortel, which lists “next-genera-tion mobility”, “enterprise trans-formation”, and “services andapplications” as the three coreelements of its strategic focus.

Alcatel will take over Nortel’sportfolio of UMTS radio-network controller (RNC), base-station (node-b) and operationadministration and maintenance(OAM) products. Completion ofthe sale is expected to occur inthe fourth quarter of 2006. Nodefinitive statements were madeby either company regarding joblosses, but Nortel stated: “It isanticipated that the significantmajority of employees of Nortel’sUMTS access business will trans-fer to Alcatel.”

Alcatel wants to buy Nortel’s UMTS business

The US operator Sprint Nextelplans to deploy a nationwideMobile WiMAX network in part-nership with Intel, Motorola andSamsung. Sprint Nextel currentlyoperates a cdma2000 EV/DO3G network and Motorola andSamsung have committed todevelop multimode devices thatoperate on both cdma2000 andMobile WiMAX.

Services will be delivered via theoperator’s licensed 2.5 GHz spec-trum and will be available to 85%

of households in the top 100 USmarkets. The operator intends toinvest about $1 bn in 2007rolling-out its network and up toan additional $2 bn in 2008.

Based on the IEEE 802.16estandard and promoted by Intel,Mobile WiMAX is expected todeliver high data rates and fullmobility via an OFDMA air inter-face. A variant of Mobile WiMAXcalled WiBRO is currently beingused in South Korea to deliverlimited commercial services.

The South Korean operator SKTelecom has joined forces withthe Chinese government todevelop TD-SCDMA technol-ogy. The two parties will establisha joint TD-SCDMA service-development centre to furtherthe commercialization of thetechnology.

Located in Bundang, Korea,the centre will focus on TD-SCDMA technology and ser-vices.

While the Chinese govern-ment is involved in a number ofTD-SCDMA joint-developmentagreements with several foreigntelecoms equipment makers, thisis its first such venture with an

outside service provider. Dubbedthe “TD-SCDMA Project”, theventure is covered by a memo-randum of understanding signedby both partners.

Through the development oftechnologies and services, SKTelecom is expected to gainaccess to China’s 3G market –once licences are finally granted.

TD-SCDMA is a 3G air inter-face that has been developed pri-marily for the Chinese market byboth Chinese and foreign com-panies. The Chinese governmentis expected to grant TD-SCDMA licences in addition tothose covering UMTS andcdma2000.

Nokia Siemens Networks hasrevealed its business plans inanticipation of the January 2007launch of the joint-venture com-pany. The company will be orga-

nized into six global businessunits called Services; RadioAccess; Service Core andApplication; Operation SupportSystems; Broadband Access; andIP Networking and Transport.

The company’s headquarterswill be in Nokia’s home town ofHelsinki, Finland, where it willbe led by Simon Beresford-Wylie,currently in charge of Nokia’sNetworks division. Four businessunits will be located in Munich,Germany, where Siemens isbased. The remaining two (RadioAccess and Operational SupportSystems) will be based in Finland.The German units will be run bycurrent Siemens executives, andthe Finnish units by currentNokia executives.

Nortel chief Mike Zafirovski admits that thecompany has fallen behind its competitorsin the 3G radio-access market.

Two months after announcingthat it will merge with ADCTelecommunications, AndrewCorporation has backed out ofthe deal, citing shareholder dis-approval with the merger.Andrew will pay ADC $10 m toexcuse itself from an agreementsigned in May 2006 and furtherpenalties could be imposed ifAndrew merges with a thirdparty within the next year.

“While we still believe in theconvergence strategy, the merger

of Andrew and ADC was onlyone method to execute againstthat,” said Ralf Faison, Andrew’schief executive.

Andrew makes RF andmicrowave equipment for a rangeof telecoms applications includ-ing cellular. The company hasalso rejected a recent unsolicitedtakeover bid from CommScope,which makes cables for the tele-coms industry. Instead, the com-pany has continued to repurchaseshares of its common stock.

Sprint Nextel backs WiMAX for 4G

Andrew merger plans quashedSouth Korean telecoms operator backsTD-SCDMA development in China

Nokia executive Simon Beresford-Wylie willlead Nokia Siemens Networks when themerged company launches in 2007.

Nokia Siemens Networks outlines plans

WEOctNews_7,8,9 5/10/06 06:40 Page 8


9B U S I N E S S

Ericsson has reorganized its cor-porate structure to reflect thegrowing convergence of telecomsservices and is recruiting 500engineers to develop next-gener-ation Internet-protocol (IP) net-works and multimedia.

Ericsson will be divided intothree business units: Networks,Global Services, and Multimedia.The Networks unit will combinethe current Systems, Access andBroadband units along withEricsson’s Power Modules andCables business interests. The unitwill develop and manufacturecore and access equipment forboth fixed and wireless networks.

It will have about 21500 employ-ees under Kurt Jofs, currentlyheading Ericsson’s Access unit.

Also new is the Multimediaunit, which will focus on thedevelopment of technologies andservices such as television, gam-ing and video. It will be formedfrom part of the Systems unitalong with the former Enterpriseunit and Ericsson MobilePlatforms and the EnterpriseLab. It will be led by Jan Wäreby,executive vice president in chargeof sales and marketing at mobile-phone maker Sony Ericsson.

The Global Services unit willnot be affected by the restructure.


Ericsson restructures

Agilent Technologies will license3D electromagnetic softwarefrom US-based Remcom for usein Agilent’s design and simula-tion software. The licence coversRemcom’s finite-difference time-domain simulation software, cur-rently sold under the name

XFDTD. The licence is notexclusive and Remcom will con-tinue to develop, sell and supportXFDTD. Agilent currently offerselectromagnetic software basedon the 3D finite-element and3D-planar method-of-momentssimulation techniques.

Revenue from the sale of cellularbase-station equipment will peakthis year and then drop sharplyuntil at least until 2011, claims areport from analyst firm In-Stat.The report “Base Station 5-YearForecast 2005–2011” also pre-dicts that the public deploymentof other wireless technologiessuch as WiFi and WiMAX willput increasing pressure on cellu-lar technologies. The report alsopoints out that the price of basestations – particularly W-CDMAequipment – has dropped signif-icantly over the past few years.

In-Stat analyst Alan Nogeesays operators are spending morethan usual on base stationsbecause they are expanding theirW-CDMA networks. “Oncedeployment of most of these newnetworks is complete, yearlyspending will drop to more typi-

cal levels,” he predicts.A report from ABI Research

suggests that the overall spend on3G equipment will exceed that of2G equipment for the first timein 2012. The report “WorldwideMobile CAPEX Investments”also predicts that the total capi-tal expenditure (CAPEX) byoperators worldwide will grow tomore than $150 bn by 2012.

ABI analyst Shailendra Pandeysays: “Mobile operators’ attitudestowards CAPEX have changedover the past two or three years.They are clearly becoming morefocused on an early return ontheir investments.” In developedmarkets, operators are focusingon improving indoor coverage aswell as the roll-out of equipmentfor the provision of advanced ser-vices such as mobile TV andmobile broadband.

Agilent licenses FDTD expertise

Base-station revenue dives, CAPEX grows

WEOctNews_7,8,9 5/10/06 06:40 Page 9


http://www.conted.ox.ac.uk/cpd/elec3

http://rakon.com

communication without bounds

UNITED STATES +1-203-630-3311 • GERMANY +49-511-676-2731 • AUSTRALIA +61-3-9751-8400 • BRAZIL +55-11-4781-2433 • CHINA +86-21-5774-4500 ©2006 Radio Frequency Systems

T h e C l e a r C h o i c e T M

RADIO FREQUENCY SYSTEMS

At Radio Frequency Systems (RFS) we’re all ears!Listening to customer’s needs and requirements.Developing and creating today’s most advanced, tailoredand cost-effective, single source, wireless solutions.

Around the world, for point-to-point or point-to-multipoint systems, operators look to RFS for the mostinnovative high performance antenna systems and theconvenience of a single source.

All backed by vast experience, an emphasis on systemsolutions, and on-site services that can includeengineering, installation and commissioning.

Radio Frequency Systems. Outstanding customerservice, and the assurance of peak wireless networkperformance.

Visit us at www.rfsworld.com

WIRELESS EUROPE 282x213mm.qxd 27/9/06 3:37 pm Page 1

http://www.rfsworld.com

1 1A N A LY S I S


Richard Jacklin looks at the challengesfacing those wanting to turn today’swireless innovation into tomorrow’sprofitable services.

My ideal working day would begin by watch-ing the BBC’s Breakfast News on my phonefrom the comfort of a seat on an express traininto London. I’m listening to the headlines instereo on my Bluetooth headset, but my bosscalls before I’ve even entered the corporateheadquarters. I’m still on my phone when Igo into the lobby, at which point my phonemakes a seamless handover to my employer’sown wireless mobile access point. And myday goes on, enhanced by a plethora of cel-lular and wireless applications and contentdesigned to enhance my life and well-being.

If you believe in such marketing hype andare convinced that today’s technology trials

will deliver a deluge of new wireless tech-nologies tomorrow, then you may be disap-pointed – just like anyone who expects a seaton a busy commuter train into London! Sowhy are we often let down by the hype andwhy does it take so long between a trial andthe eventual launch of a new technology andservice? Why do some new technologies failspectacularly – like the square steering wheel,introduced in the 1970s by British Leylandin their Austin Allegro car? What makes atechnology ubiquitous, useful and profitable?

Industry analysts use many metaphors toexplain why it is so difficult to launch a newtechnology. The development process hasbeen likened to a jigsaw puzzle, an Olympichurdles race or lining up counters in a gameof Connect 4 – essentially, any successful newtechnology must fit the bill in all key areasand enjoy some element of luck. The onlycertainty in the wireless and mobile commu-

nications industry, is that one must seethrough the hype and make a realistic assess-ment before making a financial commitmentto a new technology.

It all starts with identifying the customerneeds, wants or even fashion trends. Withoutthat, you have nothing. At around the sametime that British Leyland delivered its ludi-crous steering wheel, Renault launched ahatchback car. Renault identified a need anda gap in the market which resulted in hatch-back cars eventually being made by virtuallyevery car manufacturer in the world.

However, consumer wants and needs arenot enough and a business case must be iden-tified with regards to how money is to bemade. In the mobile communications indus-try, new ideas must be seen through the eyesof the network operators, which buy largequantities of mobile phones and distributethem through their channels. Operators oftensubsidize handsets in order to boost revenuesand if they cannot make money from a newtechnology, they will simply not buy into it.

Competing technologiesAny business case must identify exactly whowill be developing and promoting the newtechnology. There may be competing tech-nologies and large companies and consortiacan create market success by virtue of theirsize and political strength. Consider the pre-mier technology battle of the 1980s betweenthe Betamax and VHS video recording stan-dards. VHS emerged victorious after severalyears of intense competition, not because itwas clearly a superior technology, but becauseit was more attractive to the majority ofmajor equipment manufacturers.

Government regulation is another key fac-tor that must be considered when devising anew technology. An important case in pointis the UK 3G spectrum auction, which raisedsome £22.5 billion in licence fees. This isunlike more “advanced” governments aroundthe world, which decided to allocate spec-trum without a cash auction to companiesthat were going to deliver the best coverage,features and services to the population.Indeed, a progressive government shouldallocate spectrum in such a way that it bene-fits the economy as a whole.

As well as selling the spectrum, many gov-ernments also decide on which frequencybands are allocated to each technology type.Right now this is an issue for WiMAX andmobile-TV technologies such as DVB-H.Ideally those frequency bands should also beharmonized on a global basis so as to enablesimpler radio devices, greater internationalproliferation and reduced costs. Historicallygovernments can also dictate the test, certifi-

P R O D U C T D E V E L O P M E N T

Mobile successcomes aftermany hurdles

Mediaannouncement

Publicised proof-of-concept

Core specificationdevelopment

(SDO)

TechnologyDevelopment(Proprietary)

Conformantproduct

development

IOT Testspecificationdevelopment

Productvalidation at

Test Fest

Test platformand productvalidation

Host/Qualified

equipment

Test platformdevelopment

Testspecificationdevelopment

(SDO)

SDO = Standards Development OrganisationIOT = Interoperability

Test Fest = IOT Gathering

Typical cellular

product development cycle

Fig. 1. A typical product-development cyclefor a cellular product. Testing plays animportant role throughout the process.

WEOctAnal_RFI_11,12 5/10/06 16:17 Page 11


1 2 A N A LY S I S


cation and regulation of devices, although inEurope and other open market democracies,regulation plays a much smaller role. So gov-ernments can play a big part in the successor failure of any new wireless technology.

Technology maturity and price point areother factors that determine where mass-market volumes of new products can be sold.Consider the ZX-80 personal computer,which debuted in 1980 as the first computerfor £100. It was revolutionary and launched anew generation in personal computing – butit also overheated, making it a good devicefor frying eggs! A year later the next-generation ZX-81 was launched. It was morereliable, cost about £70 and more than 1 mil-lion were sold worldwide.

The mobile-phone industry is no strangerto technological maturity and price points.For example, chip suppliers had focused forsome time on delivering Bluetooth onCMOS single-chip platforms for less than£5. Such a goal is familiar to anyone involvedin today’s race for low-cost single-chip GSMtransceivers for use in mobile phones foremerging markets.

Open standards and interoperability arealso essential features of large telecom sys-

tems. GSM was built for international roam-ing through the support of inter-vendor andinter-operator connection. Now with morethan 2 billion subscribers, GSM argues thecase for standardization. However, it is oftendifficult to reach an international consensus,as illustrated by 3G licensing in China,mobile TV and fracas in the IEEE betweenultrawideband and the 802.11n standard.

Total-system-readinessTotal-system-readiness is another crucialrequirement – as illustrated by the immedi-ate and ongoing success of the i-mode mobileapplications platform and the failure of WAP.I-mode succeeded because its makers under-stood the need for all the pieces of the puzzleto be ready from day one: the network, ter-minals, applications and loads of content thatgenuinely appealed to the target audience.WAP over GPRS was dripped out bit-by-bit,and simply didn’t excite the consumer.

The final hurdle – which sometimes can bethe highest and longest in the time that mustbe invested (and relates to RFI’s business) – istest and certification. As the industry movesfrom proprietary technologies to open stan-dards, a chain of events has been set in motion

to create new core and test specifications. Thisoften involves recursive processes in which it isnecessary to update the standards when prob-lems are found. Historically, the move towardsan open standard can take two or three yearsbefore the first products are ready for high-vol-ume production. This was certainly true fortechnologies such as Bluetooth, GPRS andUMTS. Going through this process is oftenprotracted as the individual elements havetheir own meeting (or clock) rate and are notalways synchronized to the other elements asillustrated in figure 1.

So, assuming someone wants your mobiletechnology, there is money to be made, it iscompetitive, the government will help you, itworks reliably for a reasonable cost, it inter-operates with the necessary system parts, ithas been tested and everything is ready forlaunch including any content you may need– only then is there a chance that the tech-nology will be successful. You might thinkthat finding the Holy Grail is simpler, butpeople and companies do succeed and makelots of money in the process. ■

Richard Jacklin is Cellular Business Director,RFI Global Services Limited.

For Split-Second Swiss Timing Just PickYour Package and GO!

For Split-Second Swiss Timing Just Pick Your Package and GO!

These tuning-fork resonators are manufactured with a uni-que photolithographic processfor precise 32.768 kHz nominal fundamental frequencies.

Micro ceramic or metalpackage styles

SMD

High stability and low aging

Low power consumption

High shock and vibrationresistance

With ten package options, top specs and aggressive pricing, there's no need for second best.

www.microcrystal.com

Micro Crystal a Div. of ETA SAMühlestrasse 14CH-2540 GrenchenSwitzerlandTel. +41 32 655 82 82Fax +41 32 655 82 [email protected]

CC 100% Lead freeMX, MS RoHS Compliant

Benefit fromreprints ofarticles featuringyour companyReprints are invaluable sales tools. Theycan be ordered before or after publicationand can be customized to your exactneeds.

For more information and prices,contact the sales team now.

Tel: +44 (0) 117 930 1021Email: [email protected]

wireless.iop.org

Be part of your community

WEOctAnal_RFI_11,12 5/10/06 16:17 Page 12

http://www.microcrystal.com




We’ve got your destination in hand

No matter what area of WiMAX you specialize in,

Rohde & Schwarz offers the full bandwidth of test and

measurement systems you need in R&D and production:

◆ Radio conformance test systems

◆ Signal analysis products

◆ Signal generators

Get your WiMAX products to market faster.

www.rohde-schwarz.com/ad/wimax

2295_Anz_Wimax_213x282_e.indd 12295_Anz_Wimax_213x282_e.indd 1 20.09.2006 11:01:36 Uhr20.09.2006 11:01:36 Uhr

http://www.rohde-schwarz.com/ad/wimax

1 4 A N A LY S I S


By Hesham ElHamahmyWhy test mobile application performance?An industry that has made significant invest-ment to improve the quality of and access torich multimedia services can hardly affordnot to. Recent advances in the speed of

mobile data transmission and in the sophis-tication of the applications have dramaticallywidened consumers’ choice of mobile deviceapplications. The mobile device has becomea hub for communication, information andentertainment applications and is now the

“sticky” companion of the masses. To the technology or service provider,

delivering more consumer choice meansincreased network, device and applicationcomplexity. From the delay-sensitive to thevideo-intensive, today’s mobile applicationsrequire robust high-speed data pipes that relyon adaptive modulation methods that makeactual data speeds more sensitive to the RFconditions and shared network resources.Once non-issues, latency and delay are nowthe measures that subscribers use to judge thesuccess or failure of an application.

Complicating things further is the range ofdevices that rely on seamless interoperabilitybetween a greater variety of competingchipset, operating system and applicationsolutions that have produced one very com-plicated ecosystem that needs to work seam-lessly to provide a rich end-user experience.

More testingTo meet these challenges, the industry isdemanding more, repeatable testing. At6300 per handset, and a typical daily outputof 28 000 handsets, a delayed launch mightcost a manufacturer 68.4 million in lost rev-enue per day.

Ideally, application testing should supportthe entire lifecycle of an application. Initiallab testing aims to reduce deployment issuesthrough pre-launch handset verification, pre-serve operating performance through newhandset validation, and reduce field trials bygetting it right the first time. Testing using anetwork emulator, the “golden network”,means reliable and repeatable results.

In the field, applications must perform inconjunction with live content servers, billingservers and packet data networks from vari-ous sources. Here applications are testedwith a variety of drive-test solutions thatbenchmark its end-to-end performance.After deployment, ongoing monitoring isnecessary to assess how the application per-forms over time and how it copes whenother new, potentially conflicting applica-tions are deployed.

Application testing is all about maximiz-ing return on network investment andreducing churn. Today’s packet data speedsand open development environments haveopened the gates to an important new sourceof revenue for the industry. To capitalize onthat potential revenue, providers mustensure end-user satisfaction of an applica-tion’s performance – something that canonly be assured by testing. ■

Hesham ElHamahmy is senior director ofwireless business development at SpirentCommunications.

A P P L I C A T I O N S

Mobile application testing is not optional

Every masterpiece begins with the finest materials.

Mullingar Business Park, Mullingar, IrelandTel.:+353 (44) 9395600; Fax:+353 (44) 9344369

www.taconic-add.com • e-mail: [email protected]

For over 20 years, Taconic has earned the reputation of being a leader indeveloping advanced materials for microwave and high-speed digitalapplications. From our multilayer substrates to the RF line of low-cost,high-value laminates, Taconic has a RoHS-compliant solution to meetyour changing needs. Contact us today and find a better way.

Finding a better way...

WEOctAnal_Spirent_14 5/10/06 06:47 Page 14

http://www.taconic-add.com



1 5O P I N I O N


The way that handsets are marketed to the public isstifling innovation, says Graham Bell.

Handset manufacturers are good at coming up with new features. They offer their customers– the network operators – hundreds of new product concepts every year, but while some areinnovative, others are just wacky. Concept phones often represent a significant jump infunctionality from the current generation, but these new designs rarely end up in the hands ofmobile subscribers. The reason for this is the current tension between handset manufacturers– who own the brands and appeal directly to the customer – and the network operators whocontrol the sales channels.

Although leading handset makers advertise to consumers, they don’t usually sell directly tothe public. This is at odds with the fact that handset brands are strong. Nokia is a global top-ten brand, with Motorola, Panasonic and Sony also appearing in the top-100 brands,according to Interbrand Best Global Brand 2006 survey. But not one operator appears in thetop 100.

Operators heavily subsidize the handset market and ultimately decide which new productswill be available to consumers. Competition between operators is driven to some extent bytheir handset portfolios, so operators must have the best-selling models available.

There are several reasons why operators have not been eager take the risks associated withtruly innovative, revolutionary products – including the fact that the major European

networks are dominated by the conservative markets ofUK and Germany. Operators must also buy lots ofhandsets to keep the costs down, which favours productswith minimal risk. Extensive handset portfolios are atodds with limited shelf space in operators’ distributionnetworks and consumers can become confused by toomuch choice. It is also difficult for customer-supportcentres to deal with numerous handset types. While

handset brands are strong and repeat purchases high, an overwhelming proportion ofreplacement handsets are standard form factor and offer basic features. Consumers areunsophisticated when choosing a handset and a significant number of new handsets are basicphones included in pay-as-you-go packages.

Operator revenues continue to be dominated by voice and text usage and users still tend tomake buying decisions on tariffs rather than value-add services. So there is no reason foroperators to sell small volumes of fancy-featured phones that don’t generate extra revenue.Indeed, manufacturers tend to add features and enabling technologies – such as cameras andbrowsers – that increase perceived value for the consumer, but do not necessarily boostrevenues for the operator. With little differentiation between handsets and services, operatorsare looking towards developing their own handset brands such as Vodafone Live! However,working with an original design manufacturer to create a handset is costly, complex and time-consuming.

There are two ways that the situation can be improved. Firstly, software should be used as adifferentiator. Software customization can be done in smaller volumes and hence is lower riskthan more creative industrial design. Custom applications can be used to drive greater use ofthe operator’s differentiating services. Secondly, user interfaces should be improved to make iteasier for subscribers to discover and use more advanced services. A good example of a wellimplemented feature is the camera-phone, which has succeeded because it’s easy to use onmost phones – just pull back the lens cover and shoot. ■

Graham Bell is a principal consultant at PA Consulting’s Wireless Technology Group. He can becontacted at [email protected].

Innovative conceptphones rarely endup in the hands ofmobile subscribers

G R A H A M B E L L

WEOctOpinion_15 5/10/06 13:51 Page 15



http://www.ets-lindgren.com/ams8050



1 6 E D G E


Patrick Morgan explains how the integration ofGSM and EDGE transceivers is enabling high-volume handset production.

After years of development, EDGE services are now being usedon GSM networks worldwide. Over 100m EDGE handsetswere shipped in 2005, and that number is forecast to grow to350m by 2008. Eventually, built-in EDGE capability willbecome a required feature for all mid- and high-end handsets.Many network service operators have announced plans toinclude EDGE as a complement to their 3G networks, operat-ing as a fallback data service in dense urban areas or rural areaswhere full 3G coverage is not available.

EDGE is a low-cost upgrade to GSM packet data networks(GPRS) that can boost data transmission rates by nearly a factorof three to a maximum of 160 kbit/s. A recent survey by theGlobal Mobile Suppliers Association (GSA) confirms that 185operators in 96 countries are deploying GSM/EDGE, comparedwith 128 networks at the end of 2004 – 44% growth in two years.

Unlike the increasingly fragmented GSM handset market,EDGE is dominated by a few top-tier manufacturers. ManyEDGE handset manufacturers have experienced numerousmanufacturing delays, due to the production challenges of exist-ing EDGE radio designs. Radio architectures based on polarloop or polar modulation have been adversely affected due tocomplicated EDGE radio calibration requirements.

To enable high-volume production of EDGE handsets, man-ufacturers are demanding alternative radio architectures.However, the handset performance and manufacturing yield,as well as the calibration requirements, must be equivalent toGSM handsets. Components need to be multi-sourced, just likein the GSM handset market today and a handset receiver sensi-tivity of –110 dBm or better is becoming the industry standard.

To deliver these challenging requirements, some handset man-ufacturers are opting for a unique dual-transmitter architecturethat preserves the best-in-class performance for GSM with asimple, one-step calibration procedure for EDGE. Unlike con-ventional linear transmitter architectures, no transmit SAW isrequired. This architecture can be designed to be compatiblewith power amplifiers (PAs) and basebands from other vendorsin the market, allowing components to be multi-sourced.

In the polar-loop transmitter, the signal is applied to the PAthrough separate amplitude and phase feedback pathways. Polarmodulation is a variant of polar loop that operates without feed-back from the PA. In both cases, the amplitude pathway containscircuitry with a delay that must be matched very closely to the delaythrough the phase pathway to avoid serious performance degrada-tion. In a production environment, delays must be matched tovariations in process including supply voltage, frequency, outputpower, and temperature, creating a difficult design and manufac-

turing challenge. Under conditions of high PA gain, the loop maybecome unstable, causing damage to the PA or dropped calls.External loop filters complicate the production calibration byallowing unwanted parasitic coupling to the printed-circuit board(PCB), PA, and other sources of interference and noise.

Figure 1 shows a simulated 8-PSK modulation spectrum forvarious values of delay mismatch. The 3GPP standard requires amaximum spectral power of –54 dBc at a frequency offset of400 kHz. For high-volume production, handset manufacturerstypically require that the transceiver produces at least –60 dBc.With a delay mismatch of approximately 30 ns, the spectrumbecomes marginally failing.

High-precision delay matchingTo put this in perspective, the GSM network operates based ona symbol period of 3.7 µs. The tolerable amount of delay mis-match is only a fraction of a symbol period (30 ns), which neces-sitates very high precision in the delay matching calibrationprocedure for polar architectures. It also introduces extensivecoverage requirements to account for all sources of variation inthe transmitter chain.

There are other ways to simplify the calibration. As shown in fig-ure 1, a unique dual-transmitter architecture approach allows theoffset phase-locked loop (OPLL) to be modulated by the basebandI and Q signals. In this mode, the digital variable gain amplifier(DVGA) and upconversion mixer are bypassed to add minimalnoise contribution. This eliminates the need for a transmit SAWfilter. In 8-PSK mode, the OPLL is unmodulated and acts as alocal oscillator that up-converts the baseband I and Q signals.

This architecture transmits the signal amplitude and phasetogether, thereby avoiding completely the delay mismatch cali-bration required by polar architectures. It requires only a sim-ple one-step calibration procedure for EDGE to measure the

Dual-transmitter architecture

Fig. 1. Modulation spectrum with delay mismatch for polar architectures. Thetolerable amount of delay mismatch is only a fraction of a symbol period (30 ns).

–70

–80

–60

–50

–40

–30

–20

–10

0

–600 –400 –200 0 200 400 600

3GPP spec–54 dBc

Δt = 58 ns (failing spec)

0 ns(ideal 8-PSK)

29 ns

WEOctSilabs_16,17 5/10/06 16:38 Page 16


1 7E D G E

output power and a similar procedure is widely used for GSMhandsets today. By contrast, polar architectures require a mini-mum of four calibration procedures to account for IQ pre-distortion, IQ dc offsets, delay matching, and output power.

One of the stated advantages of polar architectures is its highpower-added efficiency (PAE) during 8-PSK transmission.However, handset talk time is a key measure of performance,which depends not just on PAE, but also on the average cur-rent drawn from the battery for all key blocks in the radioincluding the PA, transceiver and baseband. Talk time is directlyproportional to battery capacity and inversely proportional tothe average current drawn by the radio. Therefore the averageradio current must be minimized to obtain the longest talk time.

For EDGE handsets, the average radio current depends onwhether the transmitter is operating in GMSK or 8-PSK mode.Even for high-end EDGE handsets and PDAs, the GMSKmode will dominate transmitter usage. Speech calls, as well aslow data-rate EDGE packet data transfers are all performed inGMSK mode. The 8-PSK mode is reserved only for very high-speed data transfers, which require close proximity to the cellu-lar base station. A typical average radio current can be calculatedassuming a 90% GMSK usage mode and 10% 8-PSK usage.Despite high PA current for 8-PSK, the average radio current iscomparable to what is possible for linear and polar radios.

Receiver sensitivity is a key performance indicator for all GSMhandsets. When measured in conducted mode, a receiver sensi-tivity of –110 dBm or better is the industry standard for GMSKsignals. However, when measured with blockers present along-side the desired signal, achieving excellent receiver sensitivity ischallenging, especially for networks with EDGE services.

Figure 2 shows an example of a filtered receiver architecturecompared with a direct conversion receiver architecture. Achievingadequate AM suppression is difficult for direct-conversion

receivers. In such a direct receiver, the blocker passes through theswitch, SAW and LNA and can leak over to the local oscillator(LO) side of the mixer. If leakage is excessive, the blocker mixeswith itself and produces a dc offset at the output, corrupting thesignal. For EDGE, the blockers can be modulated in both ampli-tude and phase, preventing a simple subtraction or averaging algo-rithm for dc offset correction. Depending on the algorithmemployed, the sensitivity may become significantly degraded.

If a digital low-IF receiver is used, it prevents dc offsets fromcorrupting the signal. The dc offsets are mixed away from thesignal and filtered out by the digital filter. Hence, no dc offsetcorrection algorithm is required.

An additional concern is the integration of the synthesizer loopfilter, which is not usually integrated on direct conversion trans-ceivers. The result is that the total bill-of-materials (BOM)increases because precision low-noise components are needed, anda coupling mechanism is created for external noise sources at thePCB level to add to the phase noise of the local oscillator. If thephase noise is excessive, the blocker can mix with the phase noiseto produce additional low-frequency distortion that corrupts thereceived signal. This effect is called “reciprocal mixing” and is asignificant concern in communications systems design.

After years of development, EDGE services are a reality forglobal GSM networks. By utilizing a dual-transmitter architec-ture to preserve the best-in-class performance for GSM with asimple one-step calibration procedure for EDGE, handset man-ufacturers can reduce time to market without increasing pro-duction costs. Plus, a dual transmitter can offer the highestproven performance and be compatible with PAs and basebandsfrom all leading vendors in the market while still supportinghigh-volume production requirements for 2006 and beyond. ■

Patrick Morgan is marketing manager at Silicon Laboratories.


re brings EDGE to the massesPGA

PGA

PGA

PGA

blocker

signal

frequency

VCO + frequencysynthesizer

0/90

LNA

LNA

LNA

LNA

I/Q

1900

1800

900

850

loopfilter

signal

filteredDCoffset

signal

blocker

frequency

direct conversion

ante

nna

switc

h

PGA

VCO + frequencysynthesizer

0/90

LNA

LNA

LNA

LNA

I/Q

1900

1800

900

850 filtered receiver

ante

nna

switc

h

PGA

PGA

PGA

low IF

IADC

QADC

leakage around mixer can causeDC offsets at mixer output

base

band

anal

ogue

inte

rface

base

band

anal

ogue

inte

rface

digi

tal f

ilter

no DC offset

Fig. 2. Receiver sensitivity withblockers for filtered versus directconversion achitectures. Achievingadequate AM suppression isdifficult for direct-conversionreceivers.

WEOctSilabs_16,17 5/10/06 16:38 Page 17


1 8 O P T I M I Z A T I O N


Mobile operators are entering a new world ofliberalized radio-spectrum allocation in whichregulators are not requiring licence-holders todeploy specific services or technologies. In the-ory, liberalization will promote competitionbetween rival technologies. This is good for theradio regulator – who wants spectrum alloca-tion to be governed by market forces – andgood for the consumer – who should havemore high-quality services to choose from.

Liberalization also brings with it an “alphabetsoup” of wireless technologies that emerge andjockey for market position. For example, the2500–2690 MHz band will be offered as tech-nology-neutral spectrum. This band may appealto new and existing operators deploying a mix-ture of UMTS, HSDPA, WiMAX, mobile TVand transmission backhaul technologies.

Debate is also raging over how soon-to-be-redundant bands such as GSM at 900 MHz andanalogue broadcastTV could be re-farmed for 3Gor other services. Deploying UMTS at 900MHzoffers a more economical way to deliver 3G torural – and some dense urban – environments.

Co-exist or dieDeregulation and the trend towards technologyneutrality are making the future difficult to pre-dict. New technologies will have to co-existwith current technologies and will survive,flourish, or die in the competitive spectrummarketplace. In the US, where there is analready mature liberalized spectrum policy, multiple technolo-gies – ranging from 1G AMPS to 3.5G HSDPA – are operatedby single operators and co-exist in the same band.

A cluster of base-station sites can be configured in a near-infi-nite number of ways given the permutations of design freedomssuch as site location, height, antenna choice, sectorization,mechanical tilt, electrical tilt and azimuth. There is also a hugearray of “softer” and non-physical design parameters – such asthose associated with handover, channel powers and trafficscheduling – that influence and affect the quality of the network.

Finding the optimum or near-optimum configuration of net-

work design parameters is a key challenge fac-ing operators. The ideal configuration wouldmaximize coverage, capacity and quality-of-ser-vice while minimizing deployment costs andtime. Most GSM network design optimizationwas done using automated frequency planning(AFP) algorithms.

UMTS network design optimization toolsinvolve a tenfold increase in computationalcomplexity because coverage, capacity, qualityand interference are much more strongly cou-pled in CDMA networks. Operators and plan-ning-tool vendors are now embracing the nextgeneration of network design optimizationproducts for 3G, such as Capesso fromSymena. A good design optimization tool canboost network performance – in terms of cov-erage and capacity – by over 30% from the ini-tial manual design effort.

Different demandsAs new technologies and services are deployed,each technology layer will demand a differentoptimal configuration. This will be due to severalfactors including the fact that the air-interfaceprotocol, dynamics and operation of individualtechnologies can be completely dissimilar.

For example, GSM uses the time and fre-quency domains to achieve cell-to-cell, and chan-nel-to-channel isolation, whereas UMTS usescoding to achieve orthogonality. There are evensignificant differences between UMTS and

HSDPA, including the use of different modulation and codingschemes. In addition, HSDPA employs hard handover protocols,while UMTS uses soft. The air-interface and traffic delivery capa-bilities of different technologies result in different interference pro-tection margins, link budgets and different cell overlap tolerances.

The payload capabilities and traffic statistics associated withdifferent technologies are wide-ranging – from low-data-rate cir-cuit-switched, to bursty high-rate packet-switched connections.For example, UMTS channels may be best for “conversational”and “streaming” classes of data, whereas HSDPA is better suitedfor delivery of “interactive” and “background” classes of data.

Multiple technologies bringoptimization challenges

Liberalization of the telecoms industry is bringing new technologies onto existing base-station sites.David Barker explains how this is further complicating the network-optimization process.

FIg. 1. The optimization of multiplewireless technologies is pushing thelimits of how many different antennascan be deployed at one site.

WEOctQuintel_18,19 5/10/06 07:12 Page 18


Optimization can also be affected by the factthat each technology may have a differentrequirement for coverage – particularly if not alltechnologies are deployed at all sites. Indeed, itis likely that each technology layer will serve adifferent spatially distributed traffic population.

The location and mobility of subscribersserved by each technology will create differentinter-cell interference problems. For example,UMTS users may be served by contiguousinter-locking cells, whereas HSDPA orWiMAX users may be better served by hot-spots. If optimized independently, each tech-nology layer could demand a different ideal cellconfiguration. Of course, this is impracticaland new technology layers must operate in acooperative manner with legacy technology lay-ers. Higher order radio resource managementprocesses will aim to ensure that subscribers canmake the best use of multiple technologies.

The path to 4GThis will require the management of handoversbetween layers that are based upon trafficdemand, quality reporting, load balancingbetween layers and admission control. In manyways, the emergence of multiple technologylayers with cooperative resource management isone facet of the path to 4G, where intelligenceresides in the network. With this in mind, opti-mization of multiple technologies must con-sider not just the single-layer RF optimizationgoal in isolation, but also the expected traffictransactions between layers. This latter aspectplaces a preference that cells of different tech-nology layers should be correlated with oneanother, where handover boundaries are com-mon – a common azimuth approach.

Tilt optimization can provide the necessarycell footprint correlation between technologylayers, but the tilt-independence of differenttechnologies offers the ability to spatially filterinter-cell interference between cells of neigh-bouring sites. This is especially important whennot all technologies are deployed on all sites.

The time and cost of deployment are inte-gral and vital parameters of any optimization process. Whilenew technologies must exploit existing site infrastructure,achieving optimized technology layers would require extraantennas with different optimized tilts.

New antennas and feeder cables at existing sites to serve multi-ple technologies will inevitably place additional tower loading andresult in potential time delay and costs for any necessary site rede-velopments. There would be time and cost penalties associatedwith the planning approvals process or even planning approvalrejections. Operators may also face increased rental charges by theland and/or tower owner for additional infrastructure. Finally,there is the ever-growing public and political pressures to mini-

mize the impact that sites and site redevelop-ments have upon the environment.

Operators looking at how to optimize fordifferent technology layers within a specificband delivering different service mixes to dif-ferently distributed subscribers essentially havetwo current options to hand.

The first is to use new separate antennas –or multi-stack antennas – and new feeders.Design optimization tools can then be usedto optimize the configuration of each tech-nology layer independently. The secondoption is to use a common antenna andfeeder, and therefore a common tilt. This isoften the case on sites with constraints on theload, feeder and/or planning approval. Designtools with joint-technology optimization fea-tures can then be used to reach a comparableposition on performance.

While design optimization tools such asCapesso can support both options, one or alltechnologies will inevitably be sub-optimallyconfigured if the second option is imple-mented. To address this issue, UK-basedQuintel has developed an alternative to thesecond option.

The company is developing conventionallysized panel antennas for current and emerg-ing spectrum bands. These allow individualtechnologies to have their own beam that canbe independently tilted and optimized. Theantennas also have the ability to use existingfeeders, thereby not requiring extra feeders ata site. As such, the antennas are designed tooffer operators a predictable and reliabledeployment solution that can directly replaceexisting antennas without visual, load, orrental change, yet offer the ability to maxi-mize, optimize and manage the performanceof multiple technology layers.

Initial network planning and optimizationsimulation using Capessa has demonstrated,for example, that a 24% performance gain inHSDPA traffic served is achievable using theQuintel antennas over a common tiltapproach. The tilting technology can also be

exploited in various ways to enhance the performance of anindividual technology, rather than the overall performance ofmultiple technologies. For example, two operators could sharethe same antenna by defining two different beams or assigningdifferent cell scrambling codes at the same UMTS RF channel– effectively creating cell-splitting in the elevation plane with-out adding extra antennas. This latter technique has demon-strated better than 30% capacity increases for CDMA-basedtechnologies. ■

David Barker is Director of Technology Applications at QuintelTechnology Ltd.

Op1 Op2 Op3

service division tilt

UMTS (F1)HSDPA (F3)

tilt

Op4

UMTS-2100 band

Op1 Op2 Op3

code division tilt

UMTS (F1/SC1)UMTS (F1/SC2)

tilt

Op4

UMTS-2100 band

Op1 Op2 Op3

frequency division tilt

UMTS (F1)UNITS (F2)

tilt

Op4

UMTS-2100 band

Fig. 2. Multiple beams from one antennacan be used to optimize two differenttechnology layers (top panel). It can alsobe used to enhance the performance ofa single technology layer in several ways(centre and bottom panels).

1 9O P T I M I Z A T I O N


WEOctQuintel_18,19 5/10/06 07:12 Page 19


2 0 I M S


Avi Gabbay explains how operators shouldaddress the challenges of achieving a smoothmigration to IMS.

The IP Multimedia Subsystem (IMS) is one of themost popular topics in telecoms today and ana-lysts are predicting rapid growth. However,exactly how mobile networks and servicesmigrate towards full IMS implementation isthe subject of great debate.

According to the market-research firmInstat, the Internet Protocol (IP)Multimedia System (IMS) “will deliver the‘Holy Grail’ of convergence of access tomultimedia services/applications across anyend-user device”. There are several impor-tant factors driving the telecoms industrytowards IMS. At the top of the list are theability of IMS to reduce operational andcapital expenses; its ability to implementand introduce new services faster; and itsability to facilitate the convergence of dif-ferent core and access networks, whichallows the user to access the same servicesno matter where they are located.

Time-to-revenueThe move to IMS will allow operators tomaintain ongoing cost-reduction pro-grammes as well as protecting and growingrevenues. Aggressive competition andincreasing consumer choice in the telecomssector are making time-to-revenue a criticalfactor for success. As a result the industry isquickly turning to IMS as a way to developnew services in the fastest and cheapest way.

IMS-based platforms can be used to createnew applications that combine voice and dataand facilitate the ongoing move to fixed-mobile convergence (FMC). Indeed, IMS willmake it easier to offer just about any IP-basedservice, including voice-over-IP (VoIP), push-to-talk-over-cellular (PoC), multiparty gam-

ing, video conferencing and instant messaging – regardless ofthe location of the user and their preferred mode of access.

IMS was originally defined by the Third GenerationPartnership Project (3GPP) as part of its technology roadmapfor the UMTS-based evolution from 2G to 3G. A multimedia

domain based on the session-initiation protocol (SIP) wasthen added and support for GSM, GPRS and

WLAN was provided. Since then, 3GPP2 –which promotes the rival 3G standard

cdma2000 – has based its multimediadomain (MMD) on 3GPP’s IMS. The factthat two global 3G organizations haveembraced IMS has created a “unifying visionfor the future of telecommunications”,according to Instat.

Profound impactThe original vision for IMS was to ensurethat all Internet services – current andfuture – are accessible to users of mobilephones and other portable devices. Theimpact of IMS is profound. For example,analysts believe that IMS will boost the IP-telephony carrier market from $1.6 billion in2004, to about $4.7 billion by 2010. IMSshould also result in greater competitionamong service providers. This should benefitsubscribers with the introduction of new ser-vices, greater mobility and more specializedservices and content. As mobile devices andservices evolve, operators and equipment ven-dors are standardizing on the IMS architec-ture to create technologies that can take fulladvantage of fixed and wireless networkresources.

The many advantages of IMS have been welldocumented. These range from integratedmobility for all network applications; easier

migration of applications for fixed and mobileusers; greater quality of service; support fornew applications such as presence, videocon-ferencing, content sharing and multipartygaming; and facility for operators to deliverall these services at low levels of capital and

Overcoming challengeson the road to IMS

The Siemens CX70 is touted as the firstIMS-capable handset on the market.

Its push-and-talk facility is based on the IP Multimedia Subsystem (IMS).

WEOctIMS_20,21 5/10/06 07:04 Page 20


2 1I M S

operating expenditure. However, before the widescale adoptionof IMS can take place, the industry must overcome a long listof challenges.

A key challenge is how to migrate legacy users to the next-generation of mobile technology and services. To do this, oper-ators must test equipment and build confidence in IMS whileretaining customer loyalty through existing services. This leads

to a real dilemma for operators –what is the best way to implementIMS so that operators can providelegacy services to IMS handsets with-out investing heavily in existing,legacy platforms?

Mobile operators must recognizethat the migration to IMS will take years to complete and theirsubscribers must enjoy a smooth transition. Above all, usersmust be able to continue using trusted and familiar serviceswhile switching to new IMS devices. Operators will also find itchallenging to provide IMS services under all coverage condi-tions; to ensure IMS users enjoy seamless service continuitywhile roaming to 2G/3G coverage areas; and to introduce newIMS services to 2G/3G customers.

Providing a seamless handoff for users moving between cellu-lar and WiFi networks is one way that operators can deliver con-verged services in the run up to IMS. Cellular/WiFi handoffshould encourage the use of dual-mode handsets in any cover-age environment. Users could also make and receive calls on anIP phone, personal computer or mobile phone – using a personalprofile to manage their calls. For example, when receiving a call,there will be simultaneous ringing on all devices, one voicemailfacility and all of this will be paid for on a single prepaid account.

Operators must provide their subscribers with an IMS servicethat allows them to use current intelligent-network (IN) ser-vices – without any change to the IN service logic or platform.Again, the goal is to enable users to continue to use trusted andfamiliar services while changing to new IMS devices. On theindustry side, vendors and operators can get a head-start onlaunching IMS, extending the time in which new IMS servicescan be tested and proven to be mature enough for IMS users.

Legacy usersOperators must also enable new IMS services to be absorbed bylegacy users, easing the transition to new packet-based servicesfor those currently using circuit-switched technology. This allowsfor a wider range of subscribers to enjoy new IMS services oncurrent non-IMS handsets. This means that an operator can offerIMS services to all of its existing customers, not just IMS users.

A study conducted by Mindbranch Market Research revealedthat IMS is “not economically viable if the goal is simply toreplicate existing services in a new architecture”. The big payofffrom IMS – according to the report – will come when opera-tors use it to develop and introduce new value-added services forincremental revenue at a lower cost per subscriber. For theirpart, operators must provide for a cost-optimized infrastructurethat supports rich services – and above all allows for a smoothtransition as the new IMS architecture continues to emerge. ■

Avi Gabbay is director of product management at Outsmart.


IMS will deliverthe ‘Holy Grail’of convergence

Introducing the industry’s fastest, most flexible,and most cost-effective RF solutions.

RFwww.keithley.com

■ Faster test execution speed.

■ Quickly change test requirements with SDR architecture.

■ Unmatched ease-of-use with touch-screen GUI.

Visitwww.keithley.info/ad/RF

to learn more.

Also, please request your free copyof the booklet “Speaking RF: Wireless Terminology“ there.

DON’TWASTE YOUR

TIME, MONEY AND EFFORT

ON YESTERDAY’S RF SOLUTIONS

Belgium: [email protected] • Phone: 02-3630040Finland: [email protected] • Phone: 09-88171661

France: [email protected] • Phone: 01-64532020Germany: [email protected] • Phone: 089-849307-40

Italy: [email protected] • Phone: 02-553842Netherlands: [email protected] • Phone: 0183-635333

Sweden: [email protected] • 08-50904600 Switzerland: [email protected] • Phone: 044-8219444

United Kingdom: [email protected] • Phone: 0118-9297500

For all countries not listed contact the KIEX department in Germany:[email protected] • Phone: +49-89-84 93 07-0

The Only Available Lab Quality,Handheld RF Power Meter

RF SIGNAL GENERATOR ■ RF SIGNAL ANALYZER ■ RF POWER METER

A G R E A T E R M E A S U R E O F C O N F I D E N C E

WEOctIMS_20,21 5/10/06 07:04 Page 21

http://www.keithley.com

http://www.keithley.info/ad/RF












2 2 B A S E S T A T I O N S


Base stations have been a hotbed of development over the lastdecade. Most advances have focused on making them smaller,cheaper and more powerful. Now, with the wider acceptance ofstandard mechanical and electrical interfaces between base-sta-tion modules, a new era of development is upon us. Completelynew base-station architectures are being created and these willbring substantial benefits to telecoms operators of all types.

Today, most base-station designs comprise a standard 19-inchrack with separate modules. In the future we will see many moreoptions becoming available as the technology develops to inte-grate modules into combined units. This will lead to smaller andmore versatile architectures that are easier to site, opening upnew possibilities for base-station locations.

Some manufacturers are already exploiting increased integra-tion to build compact base-station units that can be installedby one person. These units can be deployed at sites that were notdeemed feasible for traditional equipment. Sites such as bill-boards and narrow equipment rooms can now be used to housebase stations that can provide macrocellular coverage, substan-tially reducing the operator’s site costs.

The pace of development of new types of base station is accel-erating. As well as new integration techniques and other tech-nological developments, an important catalyst has been theindustry-wide acceptance of standard mechanical and electricalinterfaces between base-station modules.

In June 2006, Open Base Station Architecture Initiative(OBSAI) announced that it had completed its initial goal ofreleasing a full set of interface, hardware and test specificationscompatible with all of the main air-interface standards in useor emerging today and making them available for public down-load. OBSAI is an industrial forum comprising more than 130module, component and base-station vendors.

The specifications enable the production of modules that fitany base station, substantially reducing the development workand costs traditionally associated with new base-station productranges. The OBSAI base-station interface specifications are com-patible with GSM, GSM/EDGE, W-CDMA, CDMA2000 andWiMAX air interfaces. Work continues with the group lookingat newer technologies, such as Long Term Evolution (LTE).

Remote RF headA rising trend in the way in which BTS technology is deployedis the remote RF head, which is fully supported by the OBSAIspecifications. A remote RF head enables network operators todeploy the active RF electronics remotely from a radio base sta-tion. All the radio-related functions are housed in a unit mountedoutside the traditional base-station cabinet and linked to an in-cabin unit that contains the digital signal processing modules.

This separation of modules makes it easier to find and appro-priate a site for the equipment and lowers the installation costs.In a typical installation the digital modules are housed in a base-ment with a cable – fibre optic, for example – running up theoutside of a building to the remote RF head on the roof or wall.Alternatively, the remote RF head can be housed at the top of anantenna mast with a fibre-optic cable linking it to the controland baseband signal-processing modules.

The remote RF head concept has some strong proponents inthe industry who see this arrangement as dominating future

base-station development. The concept certainly has many mer-its. With the mast-mounted RF head being fed by a fibre-opticlink there are no co-axial cable losses to be overcome, so a lowerpower, cheaper amplifier can be used.

However, there are disadvantages, chiefly with regard to main-tenance. A remote RF head contains complex and hot runningelectronics built into a relatively bulky unit. This can make itdifficult to maintain or replace when it is at the top of anantenna mast. Many masts in use today were also not designedto carry the concentrated mass of a remote RF head.

Maintenance requires the antenna to be shutdown and thismust be carefully coordinated if more than one operator is shar-ing the mast. Unexpected bad weather can mean maintenance isdelayed and needs to rescheduled, with the cell-site being off theair in the meantime. The resulting costs to the operator fromlost revenue can be substantial. Indeed, the business case for aremote RF head can be complex and the costs and benefits needto be carefully calculated.

Other forms of distributed base-station architecture includedistributed antenna systems in which a macrocellular base stationfeeds a grid of antennas to provide coverage to an extendedindoor area such as an airport concourse. Although these systems

A new era ofbase stationdevelopmentPeter Kenington explains how open specificationsfor the interfaces between modules is easingbase-station development.

WEOctObsai_22,23 5/10/06 07:07 Page 22


2 3B A S E S T A T I O N S

are highly cost-effective, they provide no additional capacity.A further possibility is the use of a single building to house the

control and baseband signal-processing modules for several basestations serving a wide area, such as an entire airport or a largeindustrial estate. A large number of remote RF heads can be servedfrom this single site using fibre-optic links several kilometres long.

This can be a very cost-effective arrangement. Rather than the100% redundancy desired (but often not provided) with sepa-rate base stations, redundancy can be limited to N+1 for theentire area, substantially reducing capital and operating costs.Maintenance is focused on a single site, saving operational costs.Site management is greatly simplified and rental is limited toone building, which can be located in a low rental-cost area.

The major disadvantage is the cost of the fibre-optic links. Ifexisting fibre is unavailable it would need to be installed, involv-ing considerable civil works to lay new underground cable.

There have been instances of base-station hotels beinginstalled in city centres and for major sporting events. InGlasgow, UK, for example, an arrangement is used to serve sev-eral operators in the city centre. As an historic area, planningpermission was not feasible for competing operators to installtheir own base stations and it would have been unfair to grant

permission to just one operator. A form of base-station hotelprovides the ideal solution.

The increasing need to provide indoor coverage has led tothe development of picocell base stations. Indoor coverage for3G services is particularly important with most users expectedto be accessing data services from inside buildings.

These small base stations are low cost, easy to site and are a greatway to provide indoor coverage for 3G services. In Japan, somepredictions have indicated that picocell base stations couldaccount for half the country’s base-station population and there isstrong growth also predicted in European countries and elsewhere.

Picocell base stations are ideal for providing capacity in placesin which a large number of users congregate, such as plazas andairports, railway stations or even a small company’s building. Atypical picocell base station comprises a small self-containedunit, incorporating an antenna, and is connected to a networkvia an IP link or other standard networking interface. They arecost-effective and easy to deploy.

Femtocell base stationsTaking the miniaturization concept still further is the femtocellbase station. The target for these very small and low-power basestations is to provide mobile coverage in the home or smallenterprise. The field is wide open, with most operators activelyconsidering or trialling products, although major deploymentsare not expected for about another two years.

However, the impact of the femtocell base station could befar-reaching. It’s very early days, but the widespread deploymentof domestic base stations raises several issues, not least of whichis the impact on existing macrocellular networks.

If licensing and potential interference issues are resolved, it isconceivable that widespread coverage could be achieved by allow-ing users to use other people’s home base stations when they areout and about. This could benefit operators by reducing theamount they need to invest in their public network, effectivelytheir customers would be subsidizing deployment of the publicnetwork. Femtocell base stations with HSPA capability and DSLbackhaul could even overcome the dominance of WiFi.

At the other extreme, if owners are not required to provideopen access on their femtocell base stations, the macrocellularnetwork could suffer coverage blackspots locally, effectively pre-venting nearby users from operating on the public network.This has significant regulatory implications due to the condi-tions attached to many 3G licences. With femtocell base stationscurrently under development, these issues must be resolved andpractical business models must be developed.

Base-station development is undoubtedly entering a newphase. The possibilities and options for network operators aregrowing rapidly and will enable many to enter new areas andoffer more services, more reliably for ever more users. ■

Peter Kenington is Technical Chair of the Open Base StationArchitecture Initiative (OBSAI).


Far left and bottom left: Nokia’s Flexi W-CDMA base station is small enough tobe carried by one person and can be installed in existing 2G sites. Left: The Andrew OneBase Remote Radio Head forms part of a distributedbase-station system.

WEOctObsai_22,23 5/10/06 07:07 Page 23


WEOctAdverts24 2/10/06 11:58 Page 1

http://www.narda-sts.com



http://www.spinner.de

2 5P R O D U C T F O C U S


Ran Soffer of Provigent explains howadaptive code and modulationoptimizes a backhaul link whatever thelocal conditions.

Cellular networks are evolving rapidly fromtraditional circuit-switched voice traffic tomixed voice and data traffic, which is increas-ingly reliant on packet-switched data transmis-sion. At the same time, the high-capacity dataand video services enabled by the 2.5G and 3Gradio networks have dramatically increaseddemands on the cellular backhaul network.

Bandwidth use per subscriber is growingthanks to increased competition among oper-ators to provide data services. However, therevenue generated on a per-bit, per-Hz basis isnot increasing significantly so operators areseeking efficient network technologies thatcan lower the cost of delivering data services.

Emerging technologyAdaptive code and modulation (ACM) back-haul technology provides a way forward foroperators. This emerging technology allowsoperators to achieve high-capacity data trans-mission over microwave links and improvethe link utilization. This reduces both oper-ational and capital expenditures for main-taining high-capacity links. Link modulationand forward error correction (FEC) can bechanged on the fly (depending on the linkconditions) without any loss of data. ACMcan also maintain the highest link spectralefficiency possible at any given time in anylink condition.

In traditional voice-dominated wirelessbackhaul transmission networks, serviceavailability levels of 99.999% are the norm.However, newer services such as Internetbrowsing, video streaming and video confer-encing can operate at more relaxed availabil-ity levels. ACM can separate various servicesinto different class types and then allocate therequired availability based on class. As aresult, high-class services such as voice enjoy

99.999% availability, while lower-class ser-vices like video streaming are allocated lowerlevels of availability.

ACM uses the class of service to definewhich services should be transmitted underany link condition and which services shouldbe adapted whenever the link condition isdegraded and the link payload is decreased.For example, when bad weather hasdecreased the channel capacity of a link,ACM maintains high-class services – such asE1 channels – with full bandwidth capacity

while adapting the bandwidth capacity oflow- and mid-class services such as Internetbrowsing (see figure 1).

Traffic can be mapped into different classesof service (CoS), which define the level of ser-vice for each application. Figure 2 illustrateshow different CoS – such as rich voice andvideo – are mapped into different classes ofavailability (CoA) such as 99.999% or 99.99%.

The implementation of multiple CoAincreases the available capacity up to 10 timesthat of standard links. When conditions are

ACM is the key to next-generation backhaul

BACKHAUL

time

high-classservices

low/mid-class services

256QAM 128QAM 64QAM 32QAM QPSK

256QAM

channelcapacity

99.990

99.994

99.999 rich voicespeech

video on demandpush to talk

video telephonyvideo conference

customizedinfotainmentmultimediamessaging

datatext messaging

lowCoA

mediumCoA

highCoA

rate

avai

labi

lity

Fig. 1. Channel capacity is decreased in bad weather while maintaining high-class services with full bandwidthcapacity. This is achieved by adapting the bandwidth capacity of low- and mid-class services.

Fig. 2. Different classes of service (CoS) aremapped into different classes of availability.

WEOctPFocus_Provigent_25,26 5/10/06 07:09 Page 25


2 6 P R O D U C T F O C U S


clear, the wireless link operates at maximumcapacity and provides all services with the fulldata rate. When link conditions are poor –during harsh rain, for example – predefinedhigh-availability services such as voice are notaffected. However, the capacity of lower-priority services is adapted dynamically to thechanging link conditions. This is done byprovisioning bandwidth according to the linkconditions and traffic priority.

Increasing capacityThe US-based fabless semiconductor com-pany Provigent is a pioneer in creating ACMimplementations for the wireless backhaul.The company has developed a new technol-ogy called ProviBand, which increases thecapacity of backhaul links by a factor of ten.By combining Provigent’s PVG310 single-chip modem with ProviBand, equipmentmakers can design adaptive rate systems thatcan meet the requirements of mixed TDMand IP high-capacity traffic in next-generation backhaul networks.

ACM enables changes to the forward errorcorrection (FEC) code rate and modulationon the fly according to the current conditionof the link. ACM is implemented at bothends of the link – transmitter and receiver –which ensures that changes in the modulationand/or FEC occur synchronously through-out the link. The link condition is determinedusing several indicators, thus enabling swiftadaptation of the spectral efficiency to main-tain the link during fast-fading changes.

An ACM profile defines the link param-eters (modulation and FEC) for a given rangeof the signal-to-noise ratio (SNR). The SNRrange of each profile defines the threshold forswitching from one ACM profile to another.Each ACM profile has a different spectralefficiency, derived from its modulation andFEC, as shown in figure 3. For example, theprofile labelled “High ACM Profile” in figure2 is suitable for an SNR range of 21–23 dB,and has a spectral efficiency of 6.8 bit/Hz.When link degradation causes the SNR tofall below 21 dB, a lower ACM profile is

used. An example being the “Low ACMProfile” in figure 3, which is suitable for anSNR range of 9.9–11.9 dB and delivers aspectral efficiency of 3.8 bit/Hz.

The ACM profile also defines the data rateof each service, which ensures that high-classservices will always be allocated the requiredconstant bandwidth. Bandwidth is allocated toother services based on the link condition. TheACM profile also guarantees that the total datarate of all services equals the link data rate.

The receiver continuously monitors thelink condition based on estimators like SNR,received signal level and bit-error rate. Thetrigger for switching ACM profiles variesfrom one algorithm to another. For example,it can occur when either an indicator thresh-old or an indicator gradient threshold iscrossed. Some algorithms use several indica-tors to improve the ACM switching perfor-mance and reliability. ACM algorithms willusually include hysteresis to avoid undesiredswitching when there is jitter between twoACM profiles.

Once the estimators at the receiver sideshow that the link performance is not suit-able for the current ACM profile, an ACMswitching process will be initiated. In case ofdegradation in the link performance, the newACM profile will include higher FEC and/orlower modulation, decreasing the link bit/Hzratio. The ACM switching rate is measuredin dB/s and is a key feature of ACM systems.In general, the higher the switching rate, thebetter the system’s immunity to rapid SNRchanges. When the switching is being exe-cuted, the payload rate is being modified tofit the aggregated data rate to the new avail-able link data rate.

Case studyThe following case study illustrates the ben-efits of ACM. Consider a backhaulmicrowave link operating at 23 GHz with28 MHz channel spacing and 35 dB (30 cm)antenna gain. The link is operating in a rainregion similar to Zurich, Switzerland.

The system operation was set to a mini-

mal payload of eight E1 connections plus10 Mb/s Ethernet for 99.999% availability.Using the new ACM technology, the systemwas able to operate most of the time at130 Mb/s, depending on the link conditions.Most of the time the system could support a100 Mb/s Ethernet connection instead of a10 Mb/s connection. The system automati-cally monitored the link conditions andchanged the capacity without interruptingthe data transmission (hitless changes), asshown in figure 4.

This example demonstrates how the newtechnology, based on an ACM mechanism,can play a key role in the development ofcost-effective next-generation wireless accessnetworks, by taking advantage of traffic evo-lution from synchronous TDM traffic topacket IP-based traffic.

End-to-end solutionProvigent’s complete, end-to-end ACM solu-tion supports the range of modulationschemes, from quadrature phase shift keying(QPSK) all the way to 256 quadrature ampli-tude modulation (QAM). It also manages thephysical layer and automatically changes themodulation and FEC code rate according tothe link conditions. The most suitable ACMprofile is chosen in real time – out of a widerange of available ACM profiles – in orderto reach maximum spectral efficiency in anygiven link condition. The bandwidth foreach service is adjusted automatically basedon the service’s CoS. The switching rate isvery fast, complying with a fade rate of100 dB/s to enable maintaining the link evenduring highly rapid SNR changes.

The new ACM technique enables opera-tors to support high-end 2.5/3G applications– such as video and web browsing – whilemaintaining low operational and capitalexpenditures. The technology includes all thekey potentials to move cellular networksbackhaul to new horizons and stands poisedto be one of the baseline features of any2.5/3G cellular network. www.provigent .com

BACKHAUL

8

2

1

3

4

5

6

7

spec

tral e

ffici

ency

6 8 10 12 14 16 18 20 22 24 26SNR

low ACM profile16APSKFEC=0.6

high ACM profile128QAMFEC=0.8

ProviBand

0 20

0.01

0.00140 60 80 100 120throughput (Mbps)

outa

ge p

roba

bilit

y (%

)

traditionalsystems payload

Fig. 3. Each adaptive codeand modulation (ACM)profile has a differentspectral efficiency, derivedfrom its modulation andforward error correction(FEC).

Fig. 4. In this case study,the system automaticallymonitored the linkconditions and adaptivelychanged the capacitywithout interrupting thedata.

WEOctPFocus_Provigent_25,26 5/10/06 07:09 Page 26

http://www.provigent.com


•international news coverage of pioneering developments

•features by some of the world's leading physicists

•the latest vacancies in industry and academia

•interviews with the biggest names in the physics community

Keep up to date - Subscribe to our FREE newsalert service

physicsweb.org

PWAugAdPhysicsWeb20 24/7/06 11:25 am Page 1

http://physicsweb.org

2 8 P R O D U C T F O C U S


NETWORK INTERFACES

The progress of mobile technology has beenrapid and dynamic. GSM networks that wereoriginally designed for voice have since beenupgraded to deliver data services via GPRS.But the novelty of GPRS has worn off, andnow it is perceived as being too slow by usersaccustomed to wireline broadband connec-tions. While 3G promised to support high-speed data applications – including videotelephony – its higher data rates seem paltrywhen compared to 8 Mbit/s wireline services.

Fortunately, operators have moved ondeploying the high-speed downlink packetaccess (HSDPA) enhancement to 3G.HSDPA can deliver download speeds in theMbit/s range over a true packet data connec-tion. HSDPA and its uplink counter-part (HSUPA) should satisfymost near-term dataneeds and provide3G network opera-tors with a compet-itive technology.

Cost of entryNevertheless, the costof entry for 3G opera-tors – in terms oflicence fees and roll-outcosts – has been huge,and this continues to bereflected in high charges to the user. Thereare also limited opportunities for new opera-tors to join, and in this environmentWiMAX is proving an attractive alternative,albeit a controversial one. By adopting a mul-tiple access technique that uses orthogonalfrequency channels rather than widebandspreading codes, WiMAX is touted by itssupporters as the ultimate wireless broadbandtechnology, while its detractors see it as anunnecessary addition to the wireless dataportfolio, which already includes severalcompeting technologies.

Meanwhile in the wireline industryanother clash of technologies – albeit much

more low key – has been taking place forsome time. This concerns the way that voiceand data signals are transferred reliablyaround the world. The main shift has beenfrom the original circuit-switched networksused for voice to the packet-switched net-works used for data and the Internet. Thetwo main data transfer technologies are asyn-chronous transfer mode (ATM) and, increas-ingly, Internet protocol (IP).

Wireline operators are using these technolo-gies to deliver greater bandwidths and improvetheir quality of service in this way –but mobile operators areonly just start-ing

to upgrade their network infrastructure.Backhaul interfaces must be upgraded to sup-port wireless broadband applications. Forexample, adding HSDPA to a basestation willmultiply the number of required E1/T1 back-haul links by four or more.

Choosing the right interface means bal-ancing the investment in legacy equipmentwith the need to upgrade and keep pace withthe wireline market. Other issues must alsobe considered, especially with a move to IP.End-to-end security, latency and synchro-nization are some of the main items beingaddressed by the standardization bodies.

Radio spectrum is a finite resource and

enhancements to wireless standards forincreasing the data throughput in a fixedbandwidth come at a cost. Extra costs areusually related to the installation of morebase stations or equipping existing base sta-tions with more advanced technology such asmultiple-input multiple-output (MIMO)antenna diversity.

Adequate coverageOperators must also ensure that they provide

adequate network coverage.The recent push towards

using smaller outdoor andindoor basestations will

require the use of variousbackhaul technologies ifcosts are to be mini-mized. For example, anetwork of indoorpicocells serving anoffice building

could be daisy-chainedusing an Ethernet con-

nection, with a single master basestation providing the main network access.To address these challenges, US-based

Wintegra has developed its WinPath accesspacket processor. Introduced in 2001, itsprogrammable architecture was significantlydifferent from existing devices and it was notlimited to the leading network interface stan-dards of the day. Wintegra’s cornerstonedevice is the WinComm network processor,also known as the Packet Processing Engine.This supports all on-chip data-path softwareand implements most of the layer 2 and 3protocol processing. Production-ready soft-ware for more than 50 data-path protocols issupplied, and is royalty-free with no non-recurring expenses (NRE). For wide area net-work (WAN) applications, the protocolscover ATM UNI- AAL5/AAL2, IMA,HDLC, PPP, multi-class/multi-link PPP, cir-cuit emulation and MPLS.

WinComm is a highly programmable

Mobile IP services demandgreater access speeds

Colin Alexander explains why wireless technologies must keep with pace with the rapid advances in consumer wireline data rates, which have reached 8 Mbit/s in some regions.

Fig. 1. AnAdvanced Mezzanine

Card (AMC) from SBC Technologiesthat was developed using Wintegra technology.

WEOctPFocus_Wintegra_28,29 5/10/06 16:34 Page 28


2 9P R O D U C T F O C U S


NETWORK INTERFACES

processor that is able to supportcustom protocols. It is surroundedby dedicated silicon for direct inter-facing to the most common access-equipment physical standards.Again for WAN applications, theseinclude multiple T1/E1, T3/E3,fast/gigabit Ethernet OC3/OC12ATM and POS, and VC12 chan-nelized interfaces.

This programmability and choiceof interface mean that any sup-ported protocol can be selectedindependently as a transport mech-anism on a per-port basis. Likewise,any port can switch to a new pro-tocol without any hardwarechanges. Interworking betweenprotocols is also a reality, and – astechnology moves from ATM to IP– a clear upgrade path is available.

To support the network proces-sor, WinPath includes controllers for inter-facing with external memory such as SRAM,SDRAM and flash, along with standardfunctions and interfaces for timers, inter-rupts, UART and I2C. For control-path pro-cessing, the device communicates directlywith an external PowerPC, but also includesan optional embedded MIPS communica-tions processor. This dual-processor archi-tecture removes the need for externalinterfaces, allowing WinPath to take on morethan the network access functionality. Thesoftware environment includes the WinPathDevice Driver Interface (WDDI), whichconsists of an application programminginterface (API) and a set of production-readyANSI-C drivers for easy access and full con-trol of the data-path processing. It is alsoportable and compatible with leading real-time operating systems such as VxWorks.

Figure 1 shows an Advanced MezzanineCard (AMC) platform for an A-TCA linecard that has been developed using WinPath.

The card is highly programmable, allowinga flexible and reconfigurable backhaul. Insmaller base-station applications, such asindoor picocells, WinPath can be used as partof a low-cost single-board implementation.This additional network interface flexibilityallows basestations to be daisy-chained, withWinPath used as both the master and down-stream slave interfaces.

Software functionsFrom a software perspective WinPath pro-vides classification, scheduling and manage-ment of the backhaul interface, as well as themultiplexing and de-multiplexing of multi-ple traffic streams to the baseband digital sig-nal processors (DSPs). Additional flexibilityincludes the support for media access con-trol (MAC) layer processing and schedulersfor both HSDPA/HSUPA and WiMAX.

As well as the benefits of increased inte-gration – reduced bill of materials (BOM),lower power consumption, single-board lay-

out and faster time-to-market - theprogrammability of WinPath pro-vides support for different networkaccess interfaces, as well as softwareenhancements to the wireless inter-face. These enhancements could bedriven by feedback from deployedequipment in the field, standardsmigration or the inclusion of differ-entiating product features.

If there is one certainty in thebroadband data market then it isthe relentless rise in data ratesaccompanied by increased compe-tition and customer expectation,which in turn are driving lowerprices to the consumer. The funda-mental limitations on bandwidthfor wireless communications willkeep data rates below those offeredby wireline operators but, withfalling voice revenues, there is no

shortage of opportunities available to mobileoperators as they switch their attention todata applications.

Whatever the technology, network infra-structure and the interfaces to the base sta-tions are key to their ability to handleincreased data traffic. The solution describedhere is “standards-agnostic” because the net-work interface combines an access packetprocessor with a suite of data-path protocolsimplemented in software rather than focus-ing on just one or two of the leading tech-nologies. Not only does this approach lead tolower cost and reduced power consumption,the programmability allows manufacturers tokeep pace with changing standards withouthardware upgrades, and to add their own dif-ferentiating features to achieve a potentialadvantage in a competitive market, whetherthey are addressing HSDPA/HSUPA,WiMAX or the future vision for 100 Mbit/sand beyond.www.wintegra.com

parameter memory packet memory

system features serial interfaces

serial interfacesinternal memory

MIPSmicroprocessor

non-blockingon-chipfabric

MIPS programand memory

data pathco

ntro

l pat

h

WinComm(engines)

Fig. 2. The WinPath product family has two systems that work together: the controlpath (MIPS) and the data path (WinComm) engines with traffic management.

I n n o v a t i v e S e m i c o n d u c t o r S o l u t i o n s

F o r M o b i l e a p p l i c a t i o n s

Man-Machine Interface Illumination Sensor, Power Booster, TFT Display, TFT Driver, LEDs and Drivers

Base Band and MultimediaDynastron Camera, SoC for Mobile Phones, Multi-Media Engine, System Memory, Removable Memory

Common Components

RF Discrete Components

WEOctPFocus_Wintegra_28,29 5/10/06 16:34 Page 29

http://www.wintegra.com


http://www.toshiba-components.com

3 0 P R O D U C T S

Radio network analysers arecompact and cost-effectiveRohde & Schwarz has launched a newfamily of RF instruments for testing cellularnetworks. The TSML-x products aredescribed as being compact, high-performance and cost-effective instruments.The devices can analyse signals in the80 MHz to 4 GHz frequency range, whichmakes them suitable for testing all cellular-network technologies.

The family comprises four instruments,including the TSML-CW analyser, whichperforms RF power measurements onmodulated and unmodulated carriers. Theinstrument can be triggered by time ordistance pulses and features a measurementrate of 20 channel/s for GSM signals.

Also new is the TSML-G, a GSM networkscanner that decodes system informationsuch as CI, MNC, LAC, MCC and BSIC forall GSM bands. Designed to achieve therapid optimization of GSM, GPRS andEDGE networks, the instrument can identifyroaming problems or interference issues.

The final two instruments are the

TSML-C and TSML-W PN scanners,which measure the basic RF parameters ofcdma2000 and W-CDMA networksrespectively. The devices can perform theautomatic analysis of the RF power, timingand quality parameters of all PN codes. www.rohde-schwarz.com

RF platform is for system-level testing Easy and thorough system-level testing forwireless networking and cellularcommunications systems is promised of thenew Connected Solutions Workbench fromAgilent Technologies. The Workbenchmakes RF hardware measurements such asswept error vector magnitude (EVM),complementary cumulative density function(CCDF) and bit-error-rate (BER) analysis.

According to Agilent, the Workbench canshorten the overall design cycle byeliminating the need for complex testconstruction, development of custom testsignals or data processing. The systemencapsulates tests into easy-to-usepreconfigured setups that support wireless

standards such as 3GPP, Mobile WiMAX,802.11a/b/g and ultrawideband (UWB).

The Workbench runs independently inthe PC environment and provides the high-level functions to automate the test and toperform signal creation, parameter sweeps,signal recovery, signal measurement anddata handling.www.agi lent .com

Tools support high-speed dataAndrew Corporation has introducedenhanced versions of its Invex3G networktesting tools that support the deployment ofnew high-speed data technologies bywireless operators.

Described as an industry-standard datacollection platform, Invex3G now supportsthe HSDPA enhancement of the W-CDMA3G air interface standard. Invex3G alsosupports EV-DO, an advanced datainterface for the cdma2000 3G standard.

Invex3G enables wireless networkoperators to view their services from thesubscribers’ perspective, providing criticalquality-of-service measurements. www.andrew.com


TEST ANDMEASUREMENT

The leading publication for thecompound semiconductor industry

Published monthly, each issue delivers coverage on:● Applications, R&D, business and technology news● Feature articles covering all aspects of the industry ● Opinion and analysis on the latest research and technologies● Company profiles and interviews with key industry personalities

FREE subscriptions available at compoundsemiconductor.net/subscribe

Compound Semiconductor magazine provides in-depth,timely information on current developments within thecompound semiconductor industry.

Be part of your community

WEOctProducts_30 5/10/06 11:42 Page 30

http://www.rohde-schwarz.com

http://www.agilent.com

http://www.andrew.com


http://www.compoundsemiconductor.net/subscribe

3 1T H E F U T U R E


What are the key challenges of implementingSDR in base stations? I can’t speak for the major base-station OEMs,but my understanding is that SDR technologyis evolving rapidly and being “packaged” tomeet their requirements. There is great momen-tum in the vendor community at the semi-conductor, software tools and embeddedmiddleware levels. This has resulted in thedevelopment of robust, scaleable, efficient andcost-effective SDR-enabling products.

While the OEMs appreciate that SDR will bea base-station technology, they also understandthat without robust commercial-off-the-shelf(COTS) tools and run-time technologies, it willbe too complex, risky and home-grown to yieldan acceptable return on investment. COTStools must deliver proven performance androbustness and support open architectures.Only then will SDR deliver the immediatebusiness benefits of reduced time-to-market, lower developmentcosts, cost-effective software development outsourcing andfaster product evolution.

None of these early benefits are related to the advanced featuresthat SDR will ultimately facilitate – such as multiband multi-mode systems, over-the-air reconfigurability or cognitive radio.These new features can be introduced progressively as conver-gence, regulation, security and operational issues are resolved.

Does SDR have the potential to “shake-up” theinfrastructure equipment industry? Potentially yes, because any fast-moving technology marketoffers continuous opportunity for changes in market share.Base-station production has historically been a hardware-defined business, and middle managers can be unfamiliar withsoftware-defined concepts and “shy-away” from the technologyuntil it is too late. It’s a new twist on the old battle betweenestablished and emerging technologies – in this case it’s “hard-ware-defined” versus “software-defined” engineering takendown to the embedded systems level. However, today’s marketleaders have great momentum and as long as they embrace soft-ware radio aggressively, they should be able to react to rapidtechnological evolution and new competitors.

When will we see commercial base stations with SDRtechnology?US-based Vanu offers commercial SDR base stations today. Thetop-10 cellular base-station OEMs will enter new product devel-

opment cycles starting in 2007 and are likely toadopt SDR concepts such as open architectures,componentized applications, model-drivendevelopment, and embedded middleware acrossthe GPP, DSP and FPGA.

Has the growing interest in WiMAXaccelerated the commercialization of SDR? Yes it has. But no single waveform or waveformstandard is boosting SDR more than the uncer-tainly over the proliferation of different wave-forms and waveform variants combined withthe convergence of functionality driving multi-mode, multiband systems. OEMs must be ableto get new waveforms to market as quickly aspossible, especially if they don’t even knowtoday which waveforms tomorrow will bring!Generic systems that can be configured at the“factory gate” for different air interfaces or mar-kets have a fundamental appeal to any OEM

concerned with cost and time-to-market.

What role will SDR technology play in handsets? SDR enables reconfigurability in support of ubiquitous connec-tivity. SDR will eventually allow handsets to “surf ” seamlesslyfor available spectrum and switch transparently between differ-ent air interfaces to ensure the continuous availability of serviceor to support new mobile applications. New applications willeventually include something called cognitive radio, wherebythe handset and network are “intelligent” enough to configurethemselves for best performance, based on the user’s applica-tion or quality-of-service needs. Visionaries see the day whenhandsets can even assist the user in exploiting available servicesto achieve the best result. In the more prosaic world of 2008/09,SDR will help handset makers get new products to marketfaster, with lower start-up costs and more flexible capabilities.

When will we see commercial handsets with SDR? New technologies need an insertion point that fits within thelarger systems environment. Today, 3G Long-Term Evolution(LTE) is seen as the most likely technology insertion point forSDR in consumer cellular handsets. Texas Instruments – whichis the largest supplier of DSP chips for handsets – has said thatLTE deployments will begin in 2009. SDR handsets are alreadyshipping in smaller and less demanding (in terms of power andsize) markets such as military radios. ■

Interview by Hamish Johnston, editor of Wireless Europe.

Uncertainty is driving SDRSoftware-defined radio is poised to change cellular technology, explains Steve Jennis,

senior vice president for corporate development at PrismTech.

Steve Jennis: The uncertaintysurrounding next-generation air-interfaces is driving the developmentof software-defined radio technology.

WEOctFuture_31 5/10/06 07:00 Page 31


Concerto

Concerto

Vector Fields Ltd 24 Bankside, Kidlington, Oxford, OX5 1JE, U.K. Tel: +44 (0)1865 370151 Fax: +44 (0)1865 370277 Email: [email protected]

Vector Fields Inc 1700 North Farnsworth Avenue, Aurora, IL, 60505. USATel: (630) 851 1734 Fax: (630) 851 2106 Email: [email protected]

A Cobham Group company

Fine-tuning for RF and Microwave design3D Electromagnetic Design Software

ww

w.vectorfields.com

FAST – in independent tests CONCERTOraced ahead of rival software

ACCURATE – CONCERTO combines FiniteDifference Time Domain, Method of Momentsand Finite Elements analysis so you can selectthe most accurate technique for every application

EASY TO USE – you can even pause a programme mid-analysis to free up yourcomputer for other work

FREE TO TRIAL – put CONCERTO through its paces with a free evaluation copy of the software

To order your free evaluation copy of CONCERTO, call +44 (0)1865 370151or email [email protected] or visit ourstand at European Microwave Week B548.

VF_UK_CON_WirelessEurope 20.09.06 22/9/06 9:35 am Page 1



http://www.vectorfields.com




Date post:	12-Apr-2018
Category:	Documents
Upload:	truongdiep
View:	221 times
Download:	1 times

EDGE for the masses Uncertainty drives SDR ACM …iopp.fileburst.com/we/we_01_47.pdfEDGE for the...

Documents