1. MMCN’091 School of Computing Science Simon Fraser University Bounding Switching Delay in Mobile TV Broadcast Networks Cheng-Hsin Hsu Joint Work with Mohamed Hefeeda January 19, 2008

2. MMCN’092 Motivations: TV Evolution—Mobile  Battery powered  Mobile, wireless  Small screens, lower bit rates

3. MMCN’093  Most mobile devices (phones, PDAs,...) are almost full-fledged computers  Users like to access multimedia content anywhere, anytime  Longer Prime Time viewing  More business opportunities for content providers  Market research forecasts (by 2011) -500 million subscribers, 20 billion Euros in revenue  Already deployed (or trial) networks in 40+ countries [http://www.dvb-h.org] Mobile TV: Market Demand & Potential

4. MMCN’094  Over (current, 3G) cellular networks -Third Generation Partnership Project (3GPP)  -Multimedia Broadcast/Multicast Service (MBMS) -Pros: leverage already deployed networks -Cons: Limited bandwidth (<1.5 Mb/s)  very few TV channels, low quality, and higher energy consumption for mobile devices (they work mostly in continuous mode) Mobile TV: Multiple Technologies

5. MMCN’09 Mobile TV: Multiple Technologies  Over Dedicated Broadcast Networks -T-DMB: Terrestrial Digital Media Broadcasting Started in South Korea Builds on the success of Digital Audio Broadcast (DAB) Limited bandwidth (< 1.8 Mbps) -DVB-H: Digital Video Broadcast—Handheld Extends DVB-T to support mobile devices High bandwidth (< 25 Mbps), energy saving, error protection, efficient handoff, … Open standard -MediaFLO: Media Forward Link Only Similar to DVB-H, but proprietary (Qualcomm) 5

6. MMCN’096  This is called Time Slicing -Supported (dictated) in DVB-H and MediaFLO  Need to construct Feasible Time Slicing Schemes -No receiver buffer under/over flow instances -No overlap between bursts Energy Saving for Mobile TV Receivers Time Bit Rate R r Off Burst Overhead T o

7. MMCN’097  Users usually flip through many channels  Long/variable delays are annoying  In fact, users have complained long channel switching delay on DVB-H phones -YouTube MWC 2008: Channel change comparison DVB-H vs. MediaFLO -Our experience with Nokia N92/N96 (> 5 secs)  Goal: bound maximum switching delay without sacrificing energy saving for mobile receivers Controlling Channel Switching Delay

8. MMCN’098  Switching delay has multiple components -Time slicing delay (our focus) -Frame refresh delay (till an I-frame arrives) Add more/redundant I-frames [Vadakital 07] Move I-frames closer to start of burst [Rezaei 07, 08] -Processing and Decoding delays Controlling Channel Switching Delay Time R r1 Off Burst Channel Switch Time Slicing Delay

9. MMCN’099  Reduce inter-burst periods  wastes energy  Reduce delay from 1.5 to 0.25 sec  Controlling Delay: Current Approach #1 energy saving drops from 90% to 55%

10. MMCN’0910  DVB-H standard [EN 102377, May 2007] -Suggests bundling multiple channels in one group  virtually zero switching delay within a group  But, -Delay across groups can be large -Devices receive all data of the bundle  wastes energy -How do we group channels in the first place (manual)? Controlling Delay: Current Approach #2

11. MMCN’0911  Use simulcast -Broadcast each TV channel over two burst trains -One optimized for delay (bootstrap) -The other optimized for energy saving (primary) -Devices tune to bootstrap bursts for fast playout, then tune to primary bursts for high energy saving  Systematically construct optimal time slicing schemes  Three variations -SIMU : traditional video systems (nonscalable codecs) -SIMU-S: scalable codecs -SIMU-S+: scalable codecs, bandwidth limited networks Controlling Delay: Our Approach

12. MMCN’0912  Low quality not noticed during flipping  Scalable codecs facilitate stream management  SIMU-S+ less energy saving than SIMU-S, but better bw utilization Controlling Delay: Our Approach SIMU SIMU-S+ SIMU-S

13. MMCN’09 Bounding Switching Delay 13 Our Algorithms target switching delay d m full quality rate r reduced quality rate r l time slicing scheme { }  Run at the base stations to multiplex TV channels into a traffic stream

14. MMCN’09 Time Slicing Scheme – SIMU/SIMU-S  Primary bursts:  Bootstrap bursts: 14

15. MMCN’09 Correctness and Performance – SIMU/SIMU-S  Prove the scheme is feasible  Show the scheme maximizes energy saving -First, show our scheme outperforms any scheme that does not employ simulcast idea -Then, show our scheme is optimal among all simulcast schemes  Analytically derive energy saving - for devices receiving bootstrap bursts - for devices receiving primary bursts 15

16. MMCN’09 Comparison on Energy Saving 16  SIMU-S Primary: More than 95% energy saving

17. MMCN’09 Comparison on Network Utilization  SIMU/SIMU-S incur (controllable) BW overhead  SIMU+ is BW efficient, but results in lower energy saving than SIMU/SIMU-S 17

18. MMCN’0918  Our algorithms are implemented in the IP Encapsulator Real Implementation

19. MMCN’0919 Testbed for DVB-H networks

20. MMCN’09 Experimental Setup  Implemented SIMU-S scheme in C++  Broadcast 8 TV channels for 10 min  Set the target delay to be 500 msec  Collect detailed logs that contain -time and size of each burst 20

21. MMCN’09 Experimental Setup (cont.)  Based on logs, wrote a utility to emulate a million of users  Randomly switching channels -let average watch time for each channel be 100 sec  Compute switching delay and weighted energy saving 21

22. MMCN’0922  Theoretical and empirical data match  SIMU much better than Current Analytical and Empirical Energy Saving Curves

23. MMCN’0923  SIMU-S achieves the target switching delay bound Channel Switching Delay

24. MMCN’0924  SIMU-S increases energy saving from 74% to 93% in real testbed Energy Saving

25. MMCN’0925  Studied the problem of controlling switching delay  Proposed and analyzed three optimal (in terms of energy saving) time slicing schemes  Implemented and evaluated SIMU-S in a real testbed -It met the delay bound while achieving 93% energy saving  Demo Conclusions

26. MMCN’0926 Sample Video Shot from our Testbed  Burst analysis for SIMU: 2 primary& 2 bootstrap trains