Energy-efficient Multicasting of Scalable Video Streams over WiMAX Networks

Energy-efficient Multicasting of Scalable Video Streams over WiMAX Networks

Somsubhra Sharangi, Ramesh Krishnamurti, Mohamed Hefeeda,

Senior Member, IEEE

Department of Computer Science, Simon Fraser University, Canada

IEEE Transactions on Multimedia, vol. 13, no. 1, Feb. 2011, pp. 102-115.

Outline

• Introduction

• Motivation

• Problem

• Proposed multicasting algorithm– Substream Selection Algorithm (SSA)

– Energy Efficient Substream Allocation (EESA)

• Simulation

• Conclusion

Introduction

• WiMAX supports various network services.

• One of these services is the Multicast and Broadcast Service (MBS), which can be used to deliver multimedia traffic to large-scale user communities.– Yota Telecom has recently started a mobile TV service with 25

channels over its 10 Mbps mobile WiMAX network.

– UDCast has announced plans for developing broadcast TV service supporting around 50 channels over mobile WiMAX.

Introduction

• Mobile Video Multicast/Broadcast– Mobile TV users to increase by 55% by 2015 [VisionGain10]

• Competing Technologies– LTE MBMS Low Bandwidth⇒

• WiMAX Advantage– High Bandwidth

– Better Video Quality Higher Revenue⇒

Introduction

• Multicast/Broadcast Service Data Area in Downlink Frame

Motivation

• H.264 Scalable Video Coding – Temporal, spatial and quality scalability

– Embedded stream metadata information• Supplementary Enhancement Information (SEI) Message

• Video Quality: measurement of video signal peak signal-to-noise ratio (PSNR)

Layer Data Rate (kbps) Quality (dB)

EL2 589 36.00

EL1 407 34.86

BL 170 32.0

Motivation

• Example of Scalable Videos

Stream s

Sub-stream l

Motivation

Sub-stream l EL l EL l EL l EL l EL l

… EL … EL … EL … EL … EL …

Sub-stream 5

EL 4 EL 4 EL 4 EL 4 EL 4

Sub-stream 4


Sub-stream 3


Sub-stream 2


Sub-stream 1

BL BL BL BL BL

Stream 1 Stream 2 Stream 3 Stream 4 Stream 5

Problem

• This paper focuses on optimally utilizing the WiMAX Multicast/Broadcast Service to stream multiple scalable videos to mobile receivers.– Select the optimal subset of layers from each scalable stream

– Maximize the average quality of all selected substreams

Network Environment

Network Environment

• A number of scalable video streams are available at a WiMAX base station.– Each scalable stream s, 1 s S, has at most L layers.

Name(streams)

1 Layer (sub-stream 1) 2 Layers (sub-stream 2)

Data rate r1 Quality q1 Data rate r2 Quality q2

1 187 3398 380 3615

2 548 3715 824 3845

3 466 3294 848 3468

…

…

r31, rsl: data rate of substream sl q12, qsl: PSNR of substream sl

Network Environment

• The average video quality is maximized within a scheduling window.– The Scheduling window has P frames

– Each frame can accommodate F amount of data and takes time

• Maximum amount of data that can be transmitted within the scheduling window is given as

C = PF

F F F F F F F F F

P frames

Substream Selection Algorithm (SSA)


• Let V(s, q) denote the set of substreams from stream 1, …, s– no two substreams are selected from the same stream

– total quality of the selected substreams is q.

Name(streams)



1 187 3398 380 3615

2 548 3715 824 3845

3 466 3294 848 3468

V(s, q)= {sl}={11, 22, 32}, where q=3398+3845+3468


• Let R(s, q) denote the sum of data rates selected in V(s, q)

Name(streams)



1 187 3398 380 3615

2 548 3715 824 3845

3 466 3294 848 3468

R(s, q)= 187+824+848, where q=3398+3845+3468


Name(streams)



1 187 3398 380 3615

2 548 3715 824 3845

3 466 3294 848 3468C=1500

q= 0 … 3398 … 3615 …

s1 0 187 380

s2 0

s3 0


3398

187

3615

380

3715

548

3845

824

Name(streams)



1 187 3398 380 3615

2 548 3715 824 3845

3 466 3294 848 3468C=1500

R(1,3398)=187

(total) q=3398


3398

187

3615

380

3715

548

3845

824

Name(streams)



1 187 3398 380 3615

2 548 3715 824 3845

3 466 3294 848 3468C=1500

R(1,3398)=187

(total) q=3398+3715=7113

R(1,7113)=?R(2,7713)=187+548=735

x 7113

7113

735


C=1500

3294

3398

3468

3615

3715

3845

6612

6866

6909

7009

7083

7113

7139

7183

7243

7313

7330

7460

10407

10537

10581

10624

10711

10754

10798

10928

187

380

187

380

548

824

735

1011

928

1204

466

187

848

380

548

824

653

1035

846

1014

1228

735

1290

1359

1011

1872

928

1204

1201

1477

1583

1394

1859

1679

1776

2052


C=1500

3294

3398

3468

3615

3715

3845

6612

6866

6909

7009

7083

7113

7139

7183

7243

7313

7330

7460

10407

10537

10581

10624

10711

10754

10798

10928

187

380

187

380

548

824

735

1011

928

1204

466

187

848

380

548

824

653

1035

846

1014

1228

735

1290

1359

1011

1872

928

1204

1201

1477

1583

1394

1859

1679

1776

2052


• Lower bound: Q0

• Upper bound: 2Q0

Name(streams)



1 187 3398 380 3615

2 548 3715 824 3845

3 466 3294 848 3468 =848-466=382

=3468-3294=174

/ = 0.445

Energy Efficient Substream Allocation (EESA)

Simulation Setup

• Video Encoding: H.264/SVC format

• Channel: 10 MHz

• Modulation: 16-QAM ¾

• TDD frame: 5ms

• Scheduling Window: 1 second= 200 frames

• MBS data area: 50kb

• Average bit rate of substreams: 100kbps ~ 2.5 Mbps

Simulation

• Running Time– (a) Fixed Window Size at 1s

– (b) Fixed number of streams at 20

Simulation

• Resource Utilization

Simulation

• Energy efficiency of the EESA

Simulation

Simulation

• Effect of Receiver Buffer

Conclusion

• This paper proposed energy-efficient multicasting of scalable video streams– Maximize the average video stream quality

– Reduce receiver energy consumption

TheENDThanks for your attention !

Energy-efficient Multicasting of Scalable Video Streams over WiMAX Networks

Documents