WiMAX and Mobile WiMAX

802.16-2004 (d), 802.16-2005(e)

Presentation Overview• Standard Overview • WiMAX Family

– WiMAX, Mobile WiMAX Specification Overview• Algorithm descriptions

– PHY, MAC• Mobile WiMAX performance

– Link, System, Comparative• WiMAX Availability/Deployments• Further amendments

– 802.16h, 802.16j, 802.16m

Specification Overview

802.16 Family (WiMAX)• 802.16 LOS 10-66 GHz• 802.16a 2-11 GHz (superceded by 802.16-2004)• 802.16c 2-11 GHz (superceded by 802.16-2004)• 802.16d Combined 802.16, 802.16a, 802.16c into 802.16-2004• 802.16e Approved Dec 7 2005

– Published Feb 2006• 802.16f Network Management Information Base (MIB)

– Published Dec 1, 2005• 802.16g Network management plane

– Draft Feb 2006• 802.16h Coexistence with license-exempt 802.16 protocols

– Draft• 802.16i Mobile Management Information Base (explicitly to

handle updates from 802.16e) – Just accepted contributions• 802.16j Mobile Multihop Relay (More later in presentation)• 802.16k Network Management/Bridging• 802.16m 4G WiMAX

– Just startedhttp://grouper.ieee.org/groups/802/16/milestones_active.html

IEEE 802.16 Standards

Source: www.wimaxforum.org/news/events/wimax_day_agenda/Gordon_Member_IEEE_802.16.pdf

WiMAX Schedules

WiMAX Forum (2006): Mobile WiMAX – Part I: A Technical Overview and Performance Evaluation. Available at www.wimaxforum.org

Relationship Between Players

www.wimaxforum.org

Source: www.wimaxforum.org/news/events/wimax_day_agenda/Gordon_Member_IEEE_802.16.pdf

802.16 Standard and Usage Model Mapping

802.16-2004• Actually a suite of PHY protocols

High Speed Unlicensed MAN

Source: www.wimaxforum.org/news/events/wimax_day_agenda/Gordon_Member_IEEE_802.16.pdf

WirelessMAN-SC• Single Carrier• Licensed operation• LOS only, >10 GHz• FEC = Reed-Solomon, optional Block Turbo Codes, Convolutional

Turbo Codes• Power Control• Directional antennas at subscriber units• Channel quality measurements

– RSSI– CINR

Source: www.wimaxforum.org/news/events/wimax_day_agenda/Gordon_Member_IEEE_802.16.pdf

WirelessMAN-SCa• Licensed operation• < 11 GHz• TDD and FDD duplex• TDMA uplink• Single Carrier• Variable bandwidths• Reed-Solomon and trellis

coded modulation• Optional block and

convolutional Turbo codes

• Framing for equalization, channel performance

• Robust modes for low SINR

• Space time coding transmit diversity option

• Block adaptive modulation

WirelessMAN OFDM• Licensed operation• NLOS, < 11 GHz, TDD, FDD• TDMA• OFDM modulation, 256 point

FFT– BPSK, QPSK, 16-QAM, 64-

QAM• Reed Solomon, Optional BTC,

CTC

• Uplink power control• Optional space time coding

– 2 Tx (BS), 1 RX (SS)• RSSI, CINR measurements

– Adaptive modulation• Includes Mesh Frame (optional)

Subcarriers: 192 Data, 8 Pilot, 28 Low Guard Band, 27 High Guard Band

From IEEE Std 802.16-2004

• Licensed operation• NLOS, < 11 GHz, TDD, FDD• Channel Bandwidths > 1.0 MHz, 2-

nxregulatory bandwidth• OFDM modulation, 256 point FFT

– BPSK, QPSK, 16-QAM, 64-QAM• Subchannelization

– OFDM FDMA = OFDMA– Gives flexibiltiy in channel

assignment in time and frequency• Convolutional code, Optional BTC,

CTC• Uplink power control• Optional space time coding

– 2 Tx (BS), 1 RX (SS)• RSSI, CINR measurements

– Adaptive modulation

WirelessMAN OFDMAFrom IEEE Std 802.16-2004

Wireless HUMAN

• Unlicensed operation• NLOS, < 11 GHz, TDD• Supports all PHY but 802.11SC• Adds DFS to the MAC• Defines center frequencies at 5 GHz

– 5000 + 5 nch (MHz)

From IEEE Std 802.16-2004

802.16e (Mobile WiMAX, 802.16- 2005)

• Ideally, 802.16 + mobility– Really intended for nomadic or low mobility– Not backwards compatible with 802.16-2004

• http://www.unstrung.com/document.asp?doc_ id=76862

• Approved Dec 7 2005– Published Feb 2006– http://www.ieee802.org/16/tge/schedule.html

• Direct competitor to 3G, 4G, 802.20 though WiMAX Forum says otherwise

• Receiving significant attention• Not intended for compatibility with 802.16-

2004

Scalable OFDMA

• PHY for 802.16e• Modifies OFDMA so

FFT size varies with channel bandwidth– Keeps carrier spacing

constant

• Channel update rate of 1 KHz– Channel estimation,

equalization

H. Yaghoobi, “Scalable OFDMA Physical Layer in IEEE 802.16 WirelessMAN,” Intel Technology Journal, Volume 8, Issue 3, 2004.Available online: ftp://download.intel.com/technology/itj/2004/volume08issue03/art03_scalableofdma/vol8_art03.pdf

WiMAX Forum (2006): Mobile WiMAX – Part I: A Technical Overview and Performance Evaluation. Available at www.wimaxforum.org

Mobile WiMAX Peak Rates• Block Turbo Code

and Low Density Parity Check Code (LDPC) are optional

• Convolutional Codes (CC) and Convolutional Turbo Codes must be supported

WiMAX Forum (2006): Mobile WiMAX – Part I: A Technical Overview and Performance Evaluation. Available at www.wimaxforum.org

Optional Antenna Array Support• MIMO-STC (defined

in Matrix A)• MIMO-Spatial

Multiplexing (defined by Matrix C)

• Beamforming• Operation defined by

three classes matrices for antenna different number of antennas (2x2 STC is Alamouti)

• Support for switching between approaches

• Not being deployed initially, but more later

Peak Data Rates

WiMAX Forum (2006): Mobile WiMAX – Part I: A Technical Overview and Performance Evaluation. Available at www.wimaxforum.org

MAC/Mobility Features• Frame-by-frame resource allocation• Hybrid Automatic Repeat Request (HARQ)• UL and DL Scheduling• Variable QoS• Three handoff methods

– A traditional Hard Handoff (HHO)– Fast Base Station Switching (FBSS)

• A list of reachable base stations is maintained by mobile and base stations, but base stations discard packets if not the active BS

– Macro Diversity (MDHO)• Same list is maintained, but all base stations in the list

can participate in the reception and transmission of packets.

WiBro• Korean version of 802.16e

– Phase 1 standardized by TTA of Korea (2004)– Phase 2 standardized in 2005

• Korean spectrum allocated 2002– 2.3 GHz (100 MHz)

• Harmonization 802.16e/WiBro agreed Nov 2004 – Samsung joined WiMAX Forum Dec 2004– May indicate Samsung’s guess on 4G direction

• Plans for Nationwide Korean deployment– KT & SK Telecom launched June 30, 2006 in Seoul

http://kt.co.kr/kthome/kt_info/pr/news_center/news_view.jsp?pa ge=1&no=397&gubun=1

– KT and Hanaro Telecom to jointly deploy outside of Seoul and 6 other cities http://times.hankooki.com/lpage/tech/200501/kt2005011117243 611810.htm

How does WiBRO relate to 802.16e?

• WiMAX Forum: (http://www.wimaxforum.org/news/press_releases/WiBro_and_Mobile_WiMAX_Bac kgrounder.pdf)

– “WiBro is the service name for Mobile WiMAX in Korea. WiBro uses the Mobile WiMAX System Profile. The system profile contains a comprehensive list of features that the equipment is required or allowed to support, and, as a result, WiBro offers the same capabilities and features of Mobile WiMAX.”

– It’s Mobile WiMAX, just with a different profile (frequency, bandwidth…)• Vendors: WiBRO is compatible with 802.16e, but there’s more to Mobile WiMAX

than just 802.16e compatibility and many choices in WiBRO are different from what is mandatory in 802.16e

– From (http://www.nortel.com/solutions/wimax/collateral/wimax_wibro_white_paper.pdf)• Some more important differences from white paper

– Mandatory Handoff • 802.16e = HHO• WiBRO = FBSS

– HARQ• 80.16e = Chase combine HARQ• WiBRO = Incremental redundancy HARQ

– Likely (though unclear) network layer differences

Reality on compatibility• All of these different profiles would be quite

difficult for a hardware radio to support (as the white paper points out), but…

• 802.16 is likely the first SDR standard• Leading implementation approaches appear

to be using specialized processors• Further, there exists a certification body for

interoperability (WiMAX Forum) with the first certified Mobile WiMAX products expected for the end of 2006 or the first quarter of 2007

• And a Global Roaming Alliance

Algorithm Descriptions

PHY, MAC, Security

Mandatory Convolutional Encoder in 802.16e

• Constraint length 7• Rate ½• Initialization

– OFDM mode:• Zeros encoder• Blocks padded with byte

0x00 at end– OFDMA

• Tailbiting• 6 bits appended to front,

output from last six bits of previous block discarded

– Tailbiting is slightly more bandwidth efficient (and mandatory), but much more computationally intensive

Encoder

Supported Data Rates

J. Andrews, A. Ghosh, R. Muhamed, Fundamentals of WiMAX, Prentice Hall, 2007

Optional Codes

• Optional codes:– block turbo codes, – convolutional turbo

codes,– low density parity

check (LDPC) codes

• Significant performance gains over mandatory convolutional codes without a lot of added complexity

802.16-e Turbo Encoder

Code performance

J. Andrews, A. Ghosh, R. Muhamed, Fundamentals of WiMAX, Prentice Hall, 2007

Subcarrier Permutations• Subcarrier permutation – how subcarriers

are allocated across subchannels• Burst profile – predefined combinations of

modulation, code rate and FEC type• Full Usage of Subcarriers (FUSC)

– Pilots independent, data subcarriers evenly spread out

• Downlink Partial Usage of Subcarriers– Subcarriers “randomized”– Supports segmentation and frequency

reuse factors of 1• Uplink Partial Usage of Subcarriers

– Subcarriers divided into tiles (4 subcarriers over 3 symbols)

– 8 data, 4 pilot– Good for high Doppler spread

• Tile Usage of Subcarriers (TUSC)– Downlink version of uplink PUSC

• Band Adaptive Modulation and Coding (AMC)

– All subcarriers are adjacent• Hurts frequency diversity, but simplifies

multiuser divserity• “Bins” defined as 8 data symbols plus 1 pilot

(in center of data)•J. Andrews, A. Ghosh, R. Muhamed, Fundamentals of WiMAX, Prentice Hall, 2007

Specified PHY Information• Channel Quality

Measurements– Used to adapt transmission

parameters• Modulation, coding, burst

profiles, power– Received signal strength

indicator• Mean, standard deviation

– SINR• Mean, standard deviation• Requires demodulation

• Power Control– Only directly supported on

uplink– 30 dB/s fluctuations– Should account for PAPR– MS maintains same

transmit power density (power/subcarrier)

– Maximum MS power for various modulations (backoff can vary to control PAPR)

Open Loop MIMO• Transmit diversity/space time coding

– Numerous optional schemes for 2,3,4 antennas– Most common:

• Spatial Multiplexing• Alamouti

• Frequency Hopped Diversity Code– Optional mode– First antenna transmits without modification– Second encodes over two consecutive subchannels

J. Andrews, A. Ghosh, R. Muhamed, Fundamentals of WiMAX, Prentice Hall, 2007

Closed Loop MIMO• Feedback mechanisms

–Antenna selection. The MS indicates to the BS which transmit antenna(s) should be used • Useful at highspeeds

–Antenna grouping. The MS indicates to the BS the optimum permutation of the order of the various antennas to be used with the current space/time encoding matrix

–Codebook based feedback. The MS indicates to the BS the optimum precoding matrix to be used, based on the entries of a predefined codebook.

• Sum capacity and MMSE most popular–Quantized channel feedback. The MS quantizes the MIMO channel and sends this information to

the BS, using the MIMO_FEEDBACK message.• High bandwidth, but usable in low speed environments

–Channel sounding. The BS obtains exact information about the CSI of the MS by using a dedicated and predetermined signal intended for channel sounding.

• Maximum (theoretical) capacity, maximum required bandwidth

J. Andrews, A. Ghosh, R. Muhamed, Fundamentals of WiMAX, Prentice Hall, 2007

Hybrid ARQ• HARQ – ARQ, but receiver can

use previous failed transmissions to improve estimates

• Type I HARQ– Chase combining– Retransmits until receiver gets

the packet right of failure propagates up to the network layer

• Type II HARQ– Incremental redundancy– Retransmits with successively

lower rate codes until receiver gets the packet right of failure propagates up to the network layer

Type II HARQ

J. Andrews, A. Ghosh, R. Muhamed, Fundamentals of WiMAX, Prentice Hall, 2007

MAC Convergence Sublayers• Supported Networking protocols

– ATM CS– Packet CS IPv4– Packet CS IPv6– Packet CS 802.3 (Ethernet)– Packet CS 802.1/Q VLAN– Packet CS IPv4 over 802.3– Packet CS IPv6 over 802.3– Packet CS IPv4 over 802.1/Q VLAN– Packet CS IPv6 over 802.1/Q VLAN– Packet CS 802.3 with optional VLAN tags and ROHC header

compression– Packet CS 802.3 with optional VLAN tags and ERTCP header

compression– Packet IPv4 with ROHC header compression– Packet IPv6 with ROHC header compression

Scheduling/QoS• Actual algorithms vendor specific, but 802.16e

assumes MS requests performance based off of a number of messages which the BS may or may not be able (or willing) to accommodate. – Max data flow per stream– Requested minimum data rate– Request for MBS– Maximum latency– Retransmission policy– Traffic priority (8 classes)– Tolerated Jitter

Mobile WiMAX MAC QoS Classes

WiMAX Forum (2006): Mobile WiMAX – Part I: A Technical Overview and Performance Evaluation. Available at www.wimaxforum.org

Network Entry Process

Network Entry Steps Negotiated Parameters

J. Andrews, A. Ghosh, R. Muhamed, Fundamentals of WiMAX, Prentice Hall, 2007

Other Services• Network discovery

– WiMAX supports either manual or automatic selection of networks based on user preference

– Defines protocols to support this• IP address management

– Note: packet transmission in WiMAX is based on connection identifiers instead of MAC addresses, so multicasting in IPv6 needs work

• Radio Resource Management – Mostly information management in standard, vendors can do different

things with information– Activities

• Controlling measurements by BSs and MSs• Delivering measurements to required databases • maintaining RRM databases• exchanging information between these databases within or across ASNs,• making radio resource information available to other functional entities, such

as HO control and QoS management.

802.16-2004 Security Vulnerabilities

• Replay Attack– Resend detected valid messages– Intention is to induce BS to send SS a reset

message

• AP Spoof– Subscribers are authenticated, but not

access point

• MAC Address Spoof• RNG-RSP Denial of Service

– Weaknesses in ranging (not encrypted, automatic acceptance of adaptations by SS)

• Auth Invalid Attack– “Auth Invalid” (possibly spoofed) puts

subscriber in a vulnerable state– Followed with a “Permanent Auth Reject”

message prevents all future communications until MAC reset

Nonetheless, Boom writes:“In the author’s opinion, the standard is an excellent starting point for the basis of a military tactical network. Given that the above recommendations have been applied, there would remain changes required to create a military wireless network. Because of the unique military environment and requirement for very high availability, DoD should adopt an appropriately robust spread spectrum physical layer to improve conventional jamming resistance. Second, DoD should continue to use higher layer encryption to protect end-to-end transmissions.”

Based on D. Boom, “Denial of Service Vulnerabilities in IEEE 802.16 Wireless Networks,” Thesis, Naval Post Graduate School, Sep 2004. Available online: http://www.ieee802.org/16/tge/contrib/C80216e-04_406.pdf

802.16e Security• Multiple layers

of security• Many aspects

added to address WiMAX problems

D. Pang, L. Tian, J. Hu, J. Zhou, J. Shi, “Overview and Analysis of IEEE 802.16e Security,” Available online: http://hdl.handle.net/2100/172

Security Improvements in 802.16e

• Authentication– BS identity now verified in

PKMv2• Authorization

– RSA-based authorization and EAP

– PKMv1 (2004) AAA in application layer, but in PKMv2 (802.16e) in different hierarchy

• Data confidentiality– Many more crypto

algorithms

• Data authenticity– AES CCM-Mode

• Replay attack– Some added protection,

but still vulnerable• Handoff support

– Possibly problematic– 802.16e suggests, but

does not define, pre- authorization

– Leads to key sharing between BS

D. Pang, L. Tian, J. Hu, J. Zhou, J. Shi, “Overview and Analysis of IEEE 802.16e Security,” Available online: http://hdl.handle.net/2100/172

Mobile WiMAX Performance

Effect of varying parameters on link and system performance

Link Simulation Parameters• From Chapt 11 of J.

Andrews, A. Ghosh, R. Muhamed, Fundamentals of WiMAX, Prentice Hall, 2007

• Scenarios:– AMC vs PUSC– Effect of HARQ– MIMO + Fading +

AMC– Open loop vs closed

loop– Common nonlinear

receiver structures

SISO AMC vs PUSC, PedestrianB

ER

SNR10-4

10-3

10-2

10-1

10-0

0 5 10 15 20 25 30

AMC vs PUSC: QPSK, Ped BR1/2 PUSCR1/2 AMCR3/4 PUSC R3/4 AMC

BE

R

SNR10-4

10-3

10-2

10-1

10-0

0 5 10 15 20 25 30

R1/2 PUSCR1/2 AMCR3/4 PUSC R3/4 AMC

AMC vs PUSC: 16QAM, Ped B

BE

R

SNR10-4

10-3

10-2

10-1

10-0

0 5 10 15 20 25 30

R1/2 PUSCR1/2 AMCR3/4 PUSC R3/4 AMC

BE

R

SNR10-4

10-3

10-2

10-1

10-0

0 5 10 15 20 25 30

R1/2 PUSCR1/2 AMCR3/4 PUSC R3/4 AMC

AMC vs PUSC: 16QAM, Ped AAMC vs PUSC: QPSK, Ped A

SISO AMC vs PUSC, Vehicular

BE

RSNR

10-4

10-3

10-2

10-1

10-0

0 5 10 15 20 25 30

R1/2 PUSCR1/2 AMCR3/4 PUSC R3/4 AMC

BE

R

SNR10-4

10-3

10-2

10-1

10-0

0 5 10 15 20 25 30

R1/2 PUSCR1/2 AMCR3/4 PUSC R3/4 AMC

AMC vs PUSC: QPSK, VehA30 AMC vs PUSC: QPSK, VehA120

SISO AMC vs PUSC, Summary

• AMC (modulation adaptation) outperforms PUSC (carrier adaptation) at slow speeds

• PUSC outperforms AMC at high speeds• Why?

– At Pedestrian (3 kph) coherence time is 150 ms

– At 120 kph channel coherence time reduced to 3 ms

– Feedback duration (5 ms)– At high speeds channel feedback needed

for AMC is poor predictor

• Moving from Pedestrian to Vehicular 120 causes drop in link performance

–QPSK ~1-1.5 dB–16-QAM ~ 2-2.5 QAM

• Why?–OFDM sensitive to frequency offsets

(Doppler)–Higher order modulations more sensitive to

channel estimations• Insights:

–Channel state information very important to performance

–Value to adjusting adaptation schemes based on Doppler

Effect of Channel Estimation (PUSC)

• Channel estimation via frequency domain linear minimum mean square error + partial information about channel covariance (from RMS delay spread)

• At low SNR, noise dominates• At high SNR, estimation imperfections dominate• Higher order modulation more sensitive to estimation imperfections

BE

R

SNR10-4

10-3

10-2

10-1

10-0

0 5 10 15 20 25 30

QPSK Real vs Perfect

BE

R

SNR10-4

10-3

10-2

10-1

10-0

0 5 10 15 20 25 30

16-QAM Real vs PerfectR1/2 Veh A30R1/2 Veh A120R3/4 Veh A30 R3/4 Veh A120R1/2 PerfectR3/4 Perfect

R1/2 Veh A30R1/2 Veh A120R3/4 Veh A30 R3/4 Veh A120R1/2 PerfectR3/4 Perfect

Effect of Hybrid-ARQ

• Type I HARQ = Chase Combining– All retransmissions identical to first transmission

• Type II HARQ = Incremental Redundancy– Puncture patterns vary by retransmission

Tran

smis

sion

s

SNR0

1

2

3

4

6 9 12 15 18 24 2721 30

Tran

smis

sion

s

0

1

2

3

4

SNR0 3 6 9 15 1812 21

R1/2 No HARQR1/2 HARQ IR1/2 HARQ II R3/4 No HARQR3/4 HARQ IR3/4 HARQ II

R1/2 No HARQR1/2 HARQ IR1/2 HARQ II R3/4 No HARQR3/4 HARQ IR3/4 HARQ II

HARQ, QPSK HARQ, 16-QAM

Hybrid ARQ Summary

• Benefit of HARQ is at low SINR• No Benefit at high SINR• Type II HARQ gives highest gain due to

reducing code rates

SIMO Performance and CorrelationR1/2 Ped B Varying Correlation R3/4 Ped B Varying Correlation

• QPSK, MMSE receiver, AMC

SIMO Correlation Summary

• Performance gain– 1x2 yields 3 dB gain (low SINR)– 1x4 yields 6 dB gain (low SINR)– Even more at high SINR

Effect of Correlation

AMC, R 1/2, Ped B AMC, R 3/4, Ped B

Complex correlation, ρ=0.5

PUSC, R1/2 Ped B PUSC, R3/4 Ped B

Correlated/Uncorrelated Fading Summary

• At low SNR, correlation has little effect• At higher SNR, however, the multiantenna gain is reduced by 1dB to 0.5dB, owing to the correlation in the fading

waveform. • Lower code rates are more sensitive to this correlation than are higher code rates.

• Figure 11.20 and Figure 11.21 provide link-level results for various possible open-loop and closed-loop transmit diversity schemes in WiMAX. The open-loop diversity considered here is the 2 x 2 Alamouti pace/time block cde (STBC).

• For AMC subcarrier permutation, STBC’s benefit is marginal, especially with correlated fading because STBC hardens the channel variation that band AMC is designed to exploit. On the other hand

• PUSC subcarrier permutation, as shown in Figure 11.22 and Figure 11.23 benefits significantly from 2 x 2 STBC

Gain versus SISO with AMC

Open Loop MIMO with Multiple Streams

AMC QPSK R1/2 in Ped B AMC QPSK R3/4 in Ped B

Dual Stream, Matrix B

Open Loop Diversity Summary

• Greater benefit for higher rates• Why?

– More sensitive to fades and added diversity reduces fades

Open Loop Gains over 2x2, AMC, Ped B, Dual Streams

Closed Loop Scenarios

• Antenna selection feedback (1/frame)– 3-bits specify antenna pair for each subchannel

• Codebook feedback (1/frame)– 6-bits that specify code for linear precoding for each subchannel– Code minimizes postdetection mean square error of both streams

• Quantized channel feedback– MS quantizes channels; BS chooses code as above

• Per subcarrier SVD– Optimal precoding

AMC, QPSK, R1/2 Ped B AMC, QPSK, R3/4 Ped B

Open versus Closed Summary

• Closed loop reasonably close (~1-2 dB) to each other

• However, closed loop techniques can add 5 dB in link gain over open loop techniques

Effect of Receiver Structures

• Ordered Successive Interference Cancellation (O-SIC)– SIC from highest SINR to lowest

• Maximum Likelihood Detection (MLD)– Search for most likely combination of transmitted symbols– Simplified by using MMSE followed by sphere-decoding– Optimum noniterative algorithm for MIMO receivers

PUSC, QPSK, R1/2 Ped B PUSC, QPSK, R3/4 Ped B

Receiver Structure Summary

• An iterative MAP will outperform MLD• QRM-MLD is suboptimal (and “low-complexity”) but

performs within a dB of MLD– K. Kim, J. Yue, “Joint channel estimation and detection

algorithms for MIMO OFDM,” Proceedings of Asilomar Conference of Signals, Systems, and Computers, Nov 02.

Gain over MMSE receiver at 10-4 BER 2x2 PUSC, Ped B

Link Performance Summary• Adaptive modulation lets WiMAX approach at low SINR

– High SINR limited by discrete modulation set• Turbo codes yield significant performance gain over

mandatory convolutional codes• AMC is better at low speeds, PUSC at high speeds• HARQ most effective at low SINR,

– HARQ II better than HARQ I in terms of BER• Closed-loop gives > 5 dB gain over open-loop at low

speeds (not practical at high speeds) • Advanced MIMO structures can give another 5 dB gain

System Level Performance

• From Chapt 12 of J. Andrews, A. Ghosh, R. Muhamed, Fundamentals of WiMAX, Prentice Hall, 2007

Simulation Parameters

System Configurations

• Basic assumes the BS is able to separate the two MSs using the two receive antennas.

• Enhanced configuration 1 increases the number of receive antennas in the DL from two to four thus providing higher order receive diversity in the DL, but is otherwise the same

• Enhanced configuration 2 increases the number of transmit antennas in the UL and DL

• Enhanced configuration 3 uses 4 x 2 closed-loop MIMO in the DL with antenna selection and quantized channel-feedback-based closed loop MIMO . Feedback once every 10 ms over two bands.

Basic ConfigurationPed B Ped A

Average Throughput, Ped B Average Throughput, Ped A

• Handheld assumes omni- directional antennas

• Desktop device has low-gain (3dBi-6dBi) directional antennas

• Desktop implements selection diversity from 6-8 antenna

Basic Configuration Summary

• Directionality does better, but limited benefit when already sectored• (1,1,3) is more spectrally efficient, but has poor cell-edge

performance• General tradeoff between reliability and spectral efficiency

Effect of Scheduling and Sub- carrier permutations (Handheld)

• (1,1,3) configuration• Without precoding, AMC offers limited benefit

(though still non-negligible ~14-18%)• Proportional fairness scheduler has slightly more

flexibility in exploiting multi-user diversity so sees better performance in capacity (~25%)

PUSC vs AMC PF vs Round Robin

Effect of MIMO ConfigurationsUplinkDownlink

Note: Same uplink used for basic and enhanced 1 configurations

Enhanced Profile SummaryTotal Throughput per Cell, Ped B, (1,1,3), 30 MHz

• Both receive diversity and transmit diversity improve the average throughput of a WiMAX network.

• By increasing the number of transmit antennas from two to four, the per sector throughput improves by 50 percent.

• Similarly, by increasing the number of receive antennas from two to four, the per sector throughput is increased by 80 percent

• UL throughput results do not account for the fact that a part of the UL bandwidth is used by the closed-loop MIMO feedback

Enhanced ProfilesDownlink, Ped B Downlink, Ped A

Enhance Profile Summary

• Fifth and tenth percentile DL data rates are not improved by increasing either transmit or receive diversity order.

• Implies that transmit diversity with antennas in DL is not sufficient to improve the cell-edge data rate in the case of (1,1,3) reuse

• Closed loop (4x2) however, provides significant • (1,1,3) frequency reuse will not be able to provide carrier-grade

reliability and guaranteed data rate unless closed-loop MIMO features are used.

5% & 10% Data Rate, Band AMC, Ped B (1,1,3)

System Performance Comments• Frequency planning

– (1,1,3) gives highest per-sector• But unequal distribution – cell edge performs poorly

– (1,3,3) gives good cell edge performance• But requires additional spectrum

– (1,1,3) with segmentation is a compromise solution• Scheduling algorithms w/ multi-user diversity can significantly

improve cell throughput – up to 25%• Diversity (receiver more so than transmitter) gives significant gain in

average throughput – 50-80%– Cell-edge behavior still bad enough that (1,1,3) is hard to justify– Needs closed-loop MIMO

• Closed loop MIMO gives big gains• Overall spectral efficiency (throughput/sector/total deployment

bandwidth)– Open loop MIMO 1.7 Mbps/Hz– Closed loop pedestrian 3.9 Mbps/Hz

Interference Patterns with Reuse = 1

http://www.wimaxforum.org/technol ogy/downloads/mobile_wimax_depl oyment_alternatives.pdf

Relative Capacity as function of • 19 BS, 3 sectors,

spaced 2.8 km, mix of users

• Proportional Fair scheduling

WiMAX Performance Comparisons

Technology Comparison

WiMAX Forum (2006): Mobile WiMAX – Part II: Competitive Analysis. Available at www.wimaxforum.org

Performance Comparisons with Existing 3G, 3G+

WiMAX Forum (2006): Mobile WiMAX – Part II: Competitive Analysis. Available at www.wimaxforum.org

Performance Comparisons

Mobile WiMAX: The Best Personal Broadband Experience! June 2006, Available at www.wimaxforum.org

Cost Comparison• Fewer base stations to provide same level of service with Mobile WiMAX

than HSPA or EVDO RevB• Less cost tied up in IP royalties (~2-3% vs 10-15%)• New spectrum costs• http://www.wimaxforum.org/technology/faq/

– The second generation of Subscriber Equipment is expected to be priced from $200 - $300 in 2008.

– The third-generation CPEs will be integrated into laptops and other portable devices and are expected to initially cost approximately $100 and be available in 2nd half 2008.

Mobile WiMAX: The Best Personal Broadband Experience! June 2006, Available at www.wimaxforum.org

Summary of comparisons

• Mobile WiMAX is most directly comparable not to existing 3G or 3G+ standards, but to ones coming out in a few years– LTE– EVDO, Rev C (UMB)

• Similar performance comes from similar technologies– OFDMA, MIMO, MBS, HARQ, Turbo codes, Adaptive

modulations, bandwidths, IP core, VOIP• So some convergence in technologies (at last!)

– Yet these “converged” standards have a huge number of options available, so will be difficult for a single ASIC solution

– Likely need for SDR

Mobile WiMAX Deployments and Availability

Certification, Chipsets, Products, Spectrum, Deployments

WiMAX Certification Schedule• The WiMAX Forum plans to have

five certification test labs located in the U.S., Europe, China, Korea and Taiwan by end of 2007.

– http://www.wimax.com/commentar y/news/wimax_industry_news/wim ax-forumae-designates-first-north- american-based-certification-lab

• Certification Waves– (.16-2004) Wave 1 enables a

simple air link – (.16-2004) Wave 2 adds QoS,

security, and advanced radio features for outdoor CPEs

– (.16-2004) Wave 3, adds indoor CPEs and PCMCIA cards for fixed and nomadic networks

– (.16e) Wave 4, adds hand-offs and simple mobile for 802.16e or mobile WiMax

– (.16e) Wave 5 adds full mobility

Source: http://www.wimaxforum.org/technology/downloads/ WiMAX_and_IMT_2000.pdf

WiMAX Certification Labs• AT4 Wireless

Parque Tecnologico de Andalucia Calle Severo Ochoa 2 29590 Campanillas, Málaga Spain

• Telecommunications Technology Association 267-2 Seohyun-dong Bundang-gu Seongnam-City Gyeonggi-do 463-824 Korea

• China Academy of Telecommunication Research 52 Hua Yuan Bei Lu Haidian District Beijing 100083 China

• 31 certified products• http://www.wimaxforum.org/kshow

case/view/catalog_search

Chipsets (1/4)• Beceem Communications (BCS2000)

– Wave 2, SIMO, MIMO, – baseband IC + RFIC (all PHY, MAC, RF)– 2.x and 3.x GHz bands– http://www.beceem.com/products/ms120.shtml

• Fujitsu MB86K21 SoC is Wave 2 – Wave 2 SOC– 2-11 GHz, 802.16e compliant– Either base or subscriber– http://www.fujitsu.com/downloads/MICRO/fma/pdf/wimax_mobilefs.pdf

• Sequans Communications (3 mobile products)– SQN2110

• 3 FPGA chipset for base stations• Wave 2,

– SQN1130 SOC• Baseband PHY, MAC• Wave 2

– SQN 1110• Like 1130, but for Wave 1• http://www.sequans.com/site/products.html

Chipsets (2/4)• Wavesat Umobile SOC

– Wave 2– Programmable PHY, MAC– 2x2 MIMO– Support for 802.11a/g– www.wavesat.com

• Runcom Technologies Ltd– RNA 200

• 802.16e-2005• Full PHY/MAC (no RF)• http://www.runcom.com/upload/infocenter/info_images/28012007194733RNA200%20ASIC.pdf

– Others:• RNF2000 is FPGA version• RNA2000 ASIC version

• TeleCIS Wireless– TCW 1620 Portable 802.16-2004 implementation– Lowest power SoC– PHY/MAC– MAC supports Wave 2 and Wave 3– 2x2 MIMO– Built in PCI interface

Chipsets (3/4)• Comsys Mobile

– CM1100• Mobile WiMA baseband• Wave 2 compliant• High speed support• http://www.comsysmobile.com/commaxcm1100.html

– CM1125• GSM/EDGE + Mobile WiMAX• Baseband PHY/MAC• Class-12 E/GPRS mobile• http://www.comsysmobile.com/commaxcm1125.html.

• Altair Semiconductor– ALT2150

• Mobile WiMAX SOC• Wave 2 Handset • http://www.ccpu.com/

• ApaceWave Technologies– APW-2000 SoC– Wave 2 MIMO A,B– http://www.apacewave.com/

Chipsets (4/4)

• picoChip– WiMAX Reference Designs– http://www.picochip.com/solutions/wimax– PC8520 802.16-2004 base station– PC8530 Mobile WiMAX base station– PC6530 Femtocell base station– PC8532 Wave2 Basestation PHY– All software upgradable

Intel• Chipsets still in the works, but…• Service providers already have expressed a

preference for Intel– http://www.digitimes.com/systems/a20070831PD215.

html– Nokia already ordered

• http://www.wimaxday.net/site/2007/09/27/nokia-orders-intel- chips-intel/

– Microsoft working on WiMAX drivers• http://www.wimaxday.net/site/2006/11/07/runcom-signs-deal-

with-microsoft/– Apple rumored to want Intel chips

• http://www.macnn.com/articles/05/06/01/apple.and.wimax/

More Intel• Already secured deals for use of WiMAX chips in laptops

– Hoping to repeat Centrino success– 2008 "Montevina" both Wi-Fi networking and WiMAX– Deals: Lenovo, Acer, Asus, Panasonic and Toshiba– No Deals: Dell, HP– Part of a planned “WiMAX Inside” Marketing campaign

• http://www.wimaxday.net/site/2007/08/03/intel-plans-%e2%80%9cwimax- inside%e2%80%9d-marketing

• Investing in WiMAX service providers– Bulgaria with Nexcom Bulgaria

• http://www.wimax.com/commentary/news/wimax_industry_news/intel- capital-and-mci-have-invested-in-bulgarian-wimax-operator-nexcom

– Japan with KDDI• http://www.wimax.com/commentary/news/wimax_industry_news/intel-

capital-and-mci-have-invested-in-bulgarian-wimax-operator-nexcom

More 802.16e equipment• ArrayComm

– OFDM + smart antennas– Uses IntelliCell beamforming

technology– Applied to other standards

• Navini– Adds smart antennas to

nomadic OFDM– Wireless broadband Georgia –

BellSouth August 05– Multicarrier Synchronous

Beam Forming– Adaptive modulation QPSK –

64 QAM

• Adaptix– SDR OFDMA/TDD

platform – Salvaged technology

from Broadstorm• Alvarion

– Devices support 802.16-2004 and e

Other Product Vendors

• From http://www.wimaxforum.org/kshowcase/view– Redline, Selex, Nokia-Siemens, ET Industries,

Axxcelera, Aperto Netwrosk, Alvarion, Airspan, Siemens, SR Telecom, Telsima

WiMax Frequency Allocation

http://www.wimaxforum.org/news/downloads/supercomm_2005/WF_Day_in_a_Life_with_WiMAX_Final.pdf

3.5 GHz is the international band for WiMAX

Other WiMAX Spectrum Opportunities

• 700 MHz band– http://www.xchangemag.com/articles/501/79h139171

83935.html?cntwelcome=1– Needs to support public safety in the nationwide band– No plan for WiMAX certification profile unless band

becomes global• 3G Spectrum

– Push to be included as a 3G standard• http://www.livemint.com/2007/09/06000634/India-backs-

Wimax-techon-3G-n.html

WiMAX Spectrum Alliances• Regulatory Database

– AT4 Wireless– Launched November 2006– http://www.wimaxforum.org/join/spectrum_demo/

• WiMAX Global Roaming Alliance– Brought together unlicensed providers to promote global roaming– Now defunct – Will probably come back in some form– http://www.theregister.co.uk/2006/09/29/oz_wimax_roaming_alliance/

• WiMAX Spectrum Owners' Alliance– http://www.wisoa.com/– Promotes roaming agreements– Participants:– Unwired Australia, Network Plus Mauritius, UK Broadband, Irish Broadband,

Austar Australia/Liberty Group, Telecom New Zealand, WiMAX Telecom Group, Enertel and Woosh Telecom

WiMAX TrialsM. Giles, “Wireless Broadband,” EDUCAUSE 2006, October 9, 2006

150 Fixed WiMAX trials are underway around the world.

Technology Deployment Timeline

• 2006 – Mobile WiMAX Trials– Fixed WiMAX Ramp– >150 WiMAX Trials/Networks Planned

• 2007 – Mobile WiMAX (MIMO) Trials– Mobile WiMAX (SISO) Ramp

• 2008 – Dual-Mode & Multi-Mode Handhelds– Mobile WiMAX (MIMO) Ramp

• Source: M. Giles, “Wireless Broadband,” EDUCAUSE 2006, October 9, 2006.

Clearwire Coverage

• Provides Fixed WiMAX based wireline replacement service to home + portability within coverage area

• 2 Mbps data + voice

http://www.clearwire.com/

• Founded in October 2003 by Craig O. McCaw

Fixed WiMAX for AT&T

• Fixed WiMAX services as DSL/cable competitor (like ClearWire)– Launch in 2Q 2008 for the US South in old

BellSouth spectrum– 2.3 GHz band– Already trialing system in Alaska– http://www.unstrung.com/document.asp?doc_

id=133853&f_src=unstrung_gnews

Fixed WiMAX Deployments• WiMAX I – IEEE 802.16 – 2004

– IEEE Standard issued– WiMAX Forum specification– Fixed point-to-point/point-to-multipoint– First WiMAX certified products end of year– Certified in 3.5 GHz band

• Equipment– (Intel) Alvarion Ltd., Aperto Networks Inc., Proxim Corp., Redline

Communications Inc., Siemens AG, and China's Huawei Technologies Co. and ZTE Corp

– Other: Picochip, WaveSat• Current Deployments

– $1.4 billion in revenue in 2004 (Marvedis)– Altitude (France) voice over pre-WiMAX

• Trial Deployments– Seattle - Sprint

http://www.wimaxxed.com/wimaxxed_news/sprint_motorol.html– London (2006)

http://www.wimaxxed.com/wimaxxed_news/london_councils.html– Xbox360? http://wimax.com/commentary/spotlight/wimax-xbox

Sprint XOhm• Mobile WiMAX focus of next generation

Sprint cellular network XOhm– http://www.xohm.com/latestnews.html

• Expects $2-2.5 billion in revenue by 2010, of which 80% comes from new revenue and 20% is from cannibalization. Assuming an ARPU of $30 per month ($360/year), that means 6.2 million subscribers by 2010

– http://www.wimax.com/commentary/spotligh t/zoom-on-xohm-2013-an-update-from-the- sprint-technology-summit

• Sprint device expectations– Sprint partners (Motorola, Intel,

Samsung) have said embed 50 million units by 2010 in the US. These will be mainly laptops and PDAs. Sprint alluded that it expects chipset cost to go down to $5-$15

– WiMAX incorporated in cameras and televisions, household appliances, and security systems, as well as over 50 million hand-held devices

• Fast extensive rollout planned– http://www.wimaxday.net/site/200

7/09/28/sprint-dominates-and- tantilises-wimax-world-usa-with- xohm/

– Precommercial launch in two markets by end of 2007

– Multi-market launch in early 2008

– 10,000 sites in preparation– 1750 base stations delivered

in 2007, 20,000 antennas– 2010 coverage

• 48 million homes, 4.5 million offices, 130 million consumer electronic devices

• Open network– “As long as the device is WiMAX

certified, it will work on Sprint’s network after provisioning. Sprint wants to sell services, not devices.”

More Sprint• Partnering with ClearWire to accelerate WiMAX

deployment– Roam between networks and exchange spectrum– http://www.xohm.com/news_071907.html

• Going to provide federal government connectivity via WiMAX– http://www.wimaxday.net/site/2007/06/05/sprint-plans-wimax-for-

gov%e2%80%99t-services/• Partnering with Google to provide services

– Aiming for the “digital lifestyle”– Gmail, social networking tools, location-based services and

multi-media services.– http://www.wimaxday.net/site/2007/07/27/sprint-and-google-will-

partner-for-mobile-wimax-services/

Motorola

• Focusing on being an infrastructure vendor• Lots of test trials• 2.5 GHz band in Chicago for Sprint

– http://www.wimaxday.net/site/2007/09/26/it-works- motorola-takes-wimax-out-of-the-lab-and-into-the-city/

• 25 other trials around the world– http://www.wimaxday.net/site/2007/03/28/motorola-in-

25-wimax-trials/

First WiBRO Deployment

• Nov 15, 2005 – Launch of KT’s personal broadband service

• “To prove its mobility, KT delivered two-way video, Internet and messaging broadband services, a range of devices that were located in a traveling shuttle bus that allowed conference attendees to experience mobile broadband.”

• Given Mobile WiMAX certification in late 2006.– http://www.wimaxforum.org/news/press_releases/AP

EC_release_111505_FINAL_FINAL1.pdf

Additional Deployments of Mobile WiMAX

• Arialink with Samsung products (rural Michigan)– http://www.3g.co.uk/PR/April2006/2948.htm

• Islanet (Puerto Rico)– http://www.techweb.com/wire/networking/193402237

• Wateen in Pakistan (Motorola)– http://news.com.com/2100-1039_3-6075684.html

• AT&T (Soma) in Nevada– http://telephonyonline.com/wimax/marketing/att_mobile_wimax_111606/

• Taiwan (Far Eastone)– http://www.wimaxday.net/site/2007/09/28/far-eastone-plans-wimax-roll-

out/• Telecom (Bahrain)

– http://www.gulf-daily- news.com/Story.asp?Article=194068&Sn=BUSI&IssueID=30181

Projected Market Breakdown

http://www.wisoa.net/members_logos/ecosystem-2-big.jpg

Further amendments

802.16h, 802.16j, 802.16m

802.16h• Improved Coexistence

Mechanisms for License-Exempt Operation

• Basically, a cognitive radio standard

• Incorporates many of the hot topics in cognitive radio

– Token based negotiation

– Interference avoidance

– Network collaboration– RRM databases

• Coexistence with non 802.16h systems

– Regular quiet times for other systems to transmit

From: M. Goldhamer, “Main concepts of IEEE P802.16h / D1,” Document Number: IEEE C802.16h-06/121r1, November 13-16, 2006.

General Cognitive Radio Policies in 802.16h

• Must detect and avoid radar and other higher priority systems

• All BS synchronized to a GPS clock• All BS maintain a radio environment map (not

their name) • BS form an interference community to resolve

interference differences• All BS attempt to find unoccupied channels first

before negotiating for free spectrum– Separation in frequency, then separation in time

DFS in 802.16h• Adds a generic

algorithm for performing Dynamic Frequency Selection in license exempt bands

• Moves systems onto unoccupied channels based on observations

Generic DFS Operation Figure h1(fuzziness in original)

Adaptive Channel Selection

• Used when BS turns on• First – attempt to find a

vacant channel– Passive scan– Candidate Channel

Determination– Messaging with Neighbors

• Second – attempt to coordinate for an exclusive channel

• If unable to find an empty channel, then BS attempts to join the interference community on the channel it detected the least interference

Figure h37: IEEE 802.16h-06/010 Draft IEEE Standard for Local and metropolitan area networks Part 16: Air Interface for Fixed Broadband Wireless Access Systems Amendment for Improved Coexistence Mechanisms for License-Exempt Operation, 2006-03-29

Collaboration• BS can request interfering

systems to back off transmit power

• Master BS can assign transmit timings– Intended to support up to 3

systems (Goldhammer)• Slave BS in an interference

community can “bid” for interference free times via tokens.

• Master BS can advertise spectrum for “rent” to other Master BS– Bid by tokens

• Collaboration supported via Base Station Identification Servers, messages, and RRM databases

• Interferer identification by finding power, angle of arrival, and spectral density of OFDM/OFDMA preambles

• Every BS maintains a database or RRM information which can be queried by other BS– This can also be hosted

remotely

802.16h Status

• Currently in letter ballot– Draft 2c (password protected):

http://www.ieee802.org/16/private/drafts/le/P8 0216h_D2c.zip

• Expand coverage, capacity by adding relay stations

• Intended for licensed operation• Not intended as a mesh network

– Actually a tree• Support mobile units

802.16j Mobile Multi-hop Relay• Relays controlled from base

stations• Fixed Relay

– Permanent installation– Useful for coverage holes

• Nomadic Relay– Temporary fixed installation– Extra capacity for special

events (military SDR conferences)

• Mobile Relay– Placed on mobile platform to

support users on the platform– Useful for public transport

(buses, trains)

Modified from Fig 1 in IEEE 802.16mmr-05/032

802.16j Requirements• Backwards compatible frame structure supporting both

relay frames and legacy frames• Definition of RF requirements including the relay link

frequency, duplexing and channel B/W• Relay shall support network entry for the mobile station

QoS and HARQ shall be supported by relay as defined in legacy 16e systems

• Relay supports mobile station handover• The specification shall support relay mobility• The use of multiple antennas to enhance the spectral

efficiency of the relay link• The support of more than one relay hop between MMR-

BS and MS– http://www.ieee802.org/16/relay/docs/80216j-06_016r1.pdf

802.16j Status

• Failed letter ballot 9-25-07– 67% (needed 75%)

http://ieee802.org/16/ballots/ballot28/report28. html

• Last open draft– http://www.ieee802.org/16/relay/docs/80216j-

06_026r4.zip

Mesh in 802.16-2004• Mesh protocols not fully specified in 802.16-2004

– Network Entry supported, some neighbor services– Routing? Congestion?

• 802.16-2004.3 (WirelessMAN OFDM) PHY includes frames for Mesh operation between subscribers

• SkyPilot is developing a layer for Mesh operation that sits atop 802.16-2004.3– Targeted for Public safety applications in 4.9 GHz band– http://www.wi-fiplanet.com/news/article.php/3549846

• 802.16f should aid creation of fixed mesh networks• 802.16i may help for mobile devices• If market demand exists, would likely be an

amendment that specifies mechanisms for mesh as 802.11s did for 802.11’s ad-hoc mode

802.16m• Intended to be 4G (satisfy requirements of IMT-Advanced)• http://www.ieee802.org/16/tgm/• Requirements still being defined

– http://www.ieee802.org/16/tgm/docs/80216m-07_002r1.pdf – Backwards compatible with 802.16j– Support MMR (802.16j), though not specifically part of the standard– Will define new profile WirelessMAN-OFDMA/2008 – Support interoperability with other systems– Bands under 6 GHz– Bandwidths of 5-20 MHz (others may be used based on ITU and

operator requirements)– TDD and FDD– Support MIMO and beamforming– Mobile expected to have 1 transmit and 2 receive antennas– Support E-911 services

IEEE C802.16m-07/002r1

More Draft 802.16m requirements• Minimum Peak Rate

– Downlink 6.5 bps/Hz– Uplink 2.8 bps/Hz

• Latency less than 802.16e• Radio Resource Management

– Reporting, interference management– Multicast broadcast service– “High-resolution” location determination

• Internetworking with:– 802.11 3GPP, 3GPP2

• Coverage optimized for 5 km, functional to 30-100 km• Optimized for low mobility (<15kph), maintain connection up to 350 kph• Optimized for contiguous spectrum but support discontiguous• Reuse/share bandwidth with legacy systems• Direct migration from 802.16e

IEEE C802.16m-07/002r1

802.16m Usage models

• High data rates and improved performance in legacy cell sizes

• Very high data rates in smaller cells• High mobility support• Deployment with MMR• Co-deployment with other networks• Collocation/coexistence with

PAN/LAN/WAN

WiMAX Summary

Points to Remember• Very flexible standard

– Modulation, subcarriers, coding, antenna arrays• Big performance gain from Turbo codes and closed-loop

MIMO• Different scheduling/subcarrier allocation algorithms

work better in different environments• WiMAX receiving massive commercial interest at the

moment– Sprint is way out in front with XOhm in the US– WiBRO has been doing well in Korea for a while

• Intel is a little late, but will likely dominate– “WiMAX Inside” push on laptops– Backing service providers, so they’ve doubled their bet

Points to Remember

• People are already looking towards 802.16m

• WiMAX is emerging primarily as a cellular competitor– Previously viewed as more of a backhaul

• Lots of software radio based implementations coming out– picoChip,

Useful WiMAX Resources• WiMAX Standards

– http://ieee802.org/16/pubs/80216-2004.html– http://www.ieee802.org/16/pubs/80216e.html

• Mobile System Profile Rev 1.4 (May 2007)– http://www.wimaxforum.org/technology/documents/wi

max_forum_mobile_system_profile_v1_40.pdf• WiMAX News

– http://www.ieee802.org/16/relay/docs/80216j- 06_026r4.zip

• WiMAX Forum:– http://www.wimaxforum.org/home