Next Generaton Wreless Networks: Research Challenges and Opportuntes Abhay Karandkar Professor Department of Electrcal Engneerng Indan Insttute of Technology Bombay, Mumba 400076 karand@ee.tb.ac.n Har Om ashram prert Dr Vkram Sarabha award Lecture 2010 1
Outlne Global Wreless Scenaro Indan Scenaro Key Research Issues and Technology Components of 4G Our Contrbutons to 4G Bandwdth Request Procedure n IEEE 802.16m Energy Effcent Schedulng over Wreless Channels Gong Forward: Towards Brdgng the Dgtal Dvde 2
Outlne Global Wreless Scenaro Indan Scenaro Key Research Issues and Technology Components of 4G Our Contrbutons to 4G Bandwdth Request Procedure n IEEE 802.16m Energy Effcent Schedulng over Wreless Channels Gong Forward: Towards Brdgng the Dgtal Dvde 3
Drvers for Broadband Emal VoIP Applcatons Search Plethora of servces causng Internet News traffc Open-Source Socal Communtes to grow 50% every year Peer-to-peer Audo/Vdeo/Images Real-Tme games 4
Wreless Broadband Data Rate 100x Mbps 10x Kbps 1x Mbps Ethernet IMT-A (4G) WMAX VDSL GDON ADSL2+ LTE / EVDO-Rev C (UMB) HSPA (3G) ADSL HSDPA WCDMA / EVDO GSM / EDGE Fxed Moble 5
Tmelnes for IMT-Advanced (4G) 2008 2009 2010 2011 2012 2013 Proposals Evaluaton Concensus Specfcaton Early Deployment 6
Outlne Global Wreless Scenaro Indan Scenaro Key Research Issues and Technology Components of 4G Our Contrbutons to 4G Bandwdth Request Procedure n IEEE 802.16m Energy Effcent Schedulng over Wreless Channels Gong Forward: Towards Brdgng the Dgtal Dvde 7
Scenaro n Inda Populaton Densty Map (source: www.relefweb.nt) 600 Mllon Cellphone subscrbers Second largest telecom market n the world Fewer products made n Inda large part s stll mported Hgh outflow of Foregn Currency $15 royalty outflow per handset Only country n the top telecom markets wth no umbrella body focusng on standards Low Geographcal Coverage Only 60% of Inda 25% of vllages covered Very Low Broadband penetraton 8
Outlne Global Wreless Scenaro Indan Scenaro Key Research Issues and Technology Components of 4G Our Contrbutons to 4G Bandwdth Request Procedure n IEEE 802.16m Energy Effcent Schedulng over Wreless Channels Gong Forward: Towards Brdgng the Dgtal Dvde 9
Technology Components for 4G OFDMA Mult-Antenna (MIMO) 3 2 2 1 3 3 1 3 2 Tme BS Dversty BS Beamformng Spectrum flexblty and Carrer aggregaton 20 MHz 20 MHz Aggregated carrers = 40 MHz 10
Technology Components for 4G Self organzaton and Self optmzaton Relayng relay lnk access lnk BS access lnk RS Enhanced Qualty of Servce support 11
Outlne Global Wreless Scenaro Indan Scenaro Key Research Issues and Technology Components of 4G Our Contrbutons to 4G Bandwdth Request Procedure n IEEE 802.16m Energy Effcent Schedulng over Wreless Channels Gong Forward: Towards Brdgng the Dgtal Dvde 12
Wreless Uplnk Schedulng SS1 Requests BS computes non-conflctng schedule SS2 Grants BS SS3 SS4
Bandwdth Request (BR) Procedure BR OpportuntesCurrent BR collson Grant Grant Tme Tme AMS wats for a random number of BR Opportuntes Number of BR opportuntes to wat chosen from Contenton Wndow Common channel for all AMS IEEE 802.16m Advanced Moble Staton (AMS) IEEE 802.16m Advanced Base Staton (ABS) Contenton Wndow Doubled Gold AMS Slver AMS Bronze AMS Contenton Wndow 14
Drawbacks of BR Procedure No dfferentaton No dfferentaton n wndow sze based on servce class Lacks farness New users can get channel before old users Scalng factor of 2 s used ndependent of System load Number of retres 15
Our BR Proposal (IEEE C80216m-09_1321r4) Connecton Prorty (CP) for each contendng Servce Flow CP s a functon of Servce Class & Number of Retres (BR collsons) CP parameters: Servce Class Intal / Maxmum Wndow Sze Wndow Scalng Factor CurrentBR collson Tme Grant Tme IEEE 802.16m Advanced Moble Staton (AMS) IEEE 802.16m Advanced Base Staton (ABS) Gold AMS Slver AMS Bronze AMS Contenton Wndow Changed as per CP Contenton Wndow 16
Outlne Global Wreless Scenaro Indan Scenaro Key Research Issues and Technology Components of 4G Our Contrbutons to 4G Bandwdth Request Procedure n IEEE 802.16m Energy Effcent Schedulng over Wreless Channels Gong Forward: Towards Brdgng the Dgtal Dvde 17
Wreless Channel Characterstcs Wreless Channel s characterzed by Sgnal strength varaton over tme, frequency and space Small scale varaton (Fadng) Interference Lmted battery lfe at hosts Physcal Layer no longer a fxed rate bt ppe Resource allocaton needs to take channel characterstcs nto account Sgnfcant performance gans n wreless networks by Cross-Layer Desgn 18
SNR Fluctuatons n a Multuser System User 1 SNR User 2 User 3 User 3 Tme 19
Multuser Dversty: A New Form of Dversty Channel fades ndependently for each user so Dfferent users experence dfferent channel gans Hgh probablty that some user wll have a strong channel BS schedules the user wth the strongest (best) channel Hence Opportunstc Schedulng Transmttng n favorable channel condton also mnmzes power but at the expense of delay Schedulng- Power s mnmzed subject to delay constrant 20
Energy Effcency Rate-Power relatonshp s convex Power ( P ) P 2 R 2 = 2 R 1 P 2 > 2 P 1 P 1 R 1 R 2 Rate ( R ) 21
Energy Effcent Schedulng Sngle Recever (Base Staton) and multple transmtters Base staton s the centralzed scheduler 22
Energy Effcent Schedulng Queue transton, average queue length, average power for user Problem: Mnmze the power consumpton of each user subject to delay constrant of each user Mult-objectve constraned optmzaton problem 23 n n n n n n n Q R A R I Q Q 1, 1 M n n M M Q Q 1 1 lmsup M n n n n M M R I X P P 1 1 ), ( lmsup N Q P, 1,, subject to Mnmze
Uplnk Soluton Vsualze a lnk between user and base staton as a Pontto-Pont scenaro Each user Determnes ts transmsson rate as f t was the only user Informs ths rate to the base staton The base staton schedules the user wth the hghest rate Queue transtons for a user who s scheduled and not for others Power and queue cost are approprately updated B a s e S t a t o n S c h e d u l e r ( U s e r s e l e c t o n p o l c y ) I n f o r m S c h e d u l n g D e c s o n U p d a t e P h a s e U s e r R a t e D e t e r m n t o n P h a s e Inform rate to Base Staton 24
Uplnk Soluton : Aucton Interpretaton The base staton auctons each tme slot The user quotng the hghest rate wns the bd User quote rates that are just suffcent to satsfy ther delay constrants Quotng unnecessarly hgh rates not favorable snce power mnmzaton s the objectve It can be proved that the queue lengths converge to cooperatve equlbrum and delay constrants are satsfed Ntn Salodkar, Abhay Karandkar, Vvek Borkar, A Stable On-lne algorthm for Energy Effcent Schedulng to appear n IEEE Transactons on Moble Computng 2010 25
Outlne Global Wreless Scenaro Indan Scenaro Key Research Issues and Technology Components of 4G Our Contrbutons to 4G Bandwdth Request Procedure n IEEE 802.16m Energy Effcent Schedulng over Wreless Channels Gong Forward: Towards Brdgng the Dgtal Dvde 26
Drvng Research Agenda Backhaul connectvty Low cost wreless backhaul Challenges DSL lke experence Archtecture for hgh speed Hgh speed spectral effcency at cell edge Low ARPU Low cost nfrastructure Tght ntegraton wth IP Infrastructure sharng Power Supply Very hgh energy effcency 27
Present Backhaul Scenaro Access & Devces Backhaul Core Servces TDM/ ATM/ IP / MPLS UMTS/ HSPA WF DSLAM GSM /EDGE TDM / ATM /ETHERNET Edge Router BSC/ RNC MSC Benefts Inexpensve rado (<$50) Low power (<10 W) PSTN Lcense free band Internet But desgned for Indoor Gateway Data centre Techncal Innovaton Voce Vdeo Audo Optmzed MAC for outdoor pont to pont lnk Interference management Packet for mult-hop swtchng lnk Node E1 crcut emulaton for Legacy access GSM/CDMA) 28
IP over Dstrbuted Cellular Archtecture Present Scenaro - Herarchcal Wreless access connect to packet core Heavy Lnk Layer Handover and QoS at access Network Dscovery and topology between access and core IP attachment at core 29
IP over Dstrbuted Cellular Archtecture TDM/ ATM/ IP / MPLS Wreless access network as IP network of base statons Self Confgurng Dstrbuted Archtecture TDM/ ATM/ IP / MPLS IP IP Base staton cooperate to manage topology dscovery Moblty, QoS and network dscovery managed at IP layer 30
Concludng Remarks Inda one of the largest telecom market Indan requrements and IPR must get reflected nto nternatonal standards Sgnfcant opportunty to push our research nto next generaton wreless networks Ths wll catalyze manufacturng scenaro n Inda We have outlned representatve efforts Bandwdth Request procedures for QoS n 4G Energy Effcent Schedulng More efforts needed- academa and research labs need to address relevant research problems 31
Acknowledgements Jont work wth Prateek Kapada, Gaur Josh, B Srnadh, Dr Ntn Salodkar. Prof Bhaskaran Raman (IITB), Prof Vvek Borkar (TIFR) Fundng support from TTSL-IIT Bombay Center for Excellence n Telecom (TICET) 32