A Network-Aware approach for Video and Metadata Streaming

1. A Network-Aware approach for Video and Metadata Streaming Authors: Aravindan Raghuveer, Member, IEEE, Ewa Kusmierek David. H. C. Du, Fellow, IEEE Presented By: Snigdha Potharaju Tanushree Kumar

2. Introduction and Motivation • Unified Architecture for video and metadata delivery • Principles of Network Aware Adaptation • Dynamic Rate Adaptation • Evaluation and Experimental Results • Control data Compression and Delivery • Evaluation and Results • Related Work • Conclusion Overview

3. Multimedia applications have stringent timing requirements for data consumption. • Bandwidth, Delay, Loss Demands on the underlying network • But, Internet is a BEST effort network Introduction

4. In order to counter the challenges posed by the Internet in the streaming of video a network aware demand adaptation technique is used. • Server -> Varies its sending rate to adapt to the network and client conditions • Client -> Contains the adaption logic Motivation

5. Fig. 1. Unified architecture for video and metadata delivery.

6. Network Status Evaluation • Available Bandwidth: If Ra < Rs , then Ba < Rs If Ra >= Rs , then Ba >= Rs • Network Backlog MECN – Multilevel ECN Principles of Network Aware Adaptation

7. Principles of Network Aware Adaptation

8. Algorithm for dynamic rate increase • Algorithm for dynamic rate decrease • Algorithm for quality increase • Dynamic Rate(D-Rate), Quality Adaption • Network Resource availability metrics Arrival rate Network Backlog • Client metric :Buffer Occupancy (convertToFrames(data)) • TCP Friendliness of D-Rate Dynamic Rate Adaptation

9. convert to frames

10. Simulation Testbed • Video Server and client • MECN capable routers • Cross Traffic generators Evaluation and Experimental Results

11. Scenario -1 (Video Trace for MPEG-4 encoded Star Wars is used) Evaluation and Experimental Results

12. Evaluation and Experimental Results

13. Evaluation and Experimental Results

14. Evaluation and Experimental Results

15. Evaluation and Experimental Results

16. Evaluation and Experimental Results

17. Evaluation and Experimental Results

18. Evaluation and Experimental Results

19. Evaluation and Experimental Results

20. Evaluation and Experimental Results

21. Evaluation and Experimental Results

22. Control data compression • Client algorithm for the reduced trace Control Data Compression and Delivery

23. Evaluation and Results

24. Evaluation and Results

25. Evaluation and Results

26. Evaluation and Results

27. Evaluation and Results

28. Delivery of control data • Essential properties • DART : A Dynamic Scheduling Algorithm for Reduced Trace Delivery Control Data Compression and Delivery

29. Evaluation and Results

30. Evaluation and Results

31. Network aware adaptation techniques • Receiver driven In the receiver driven technique the client modifies its bandwidth demands based on network status. The server transmits a base layer and multiple enhancement layers of a video over unique multicast channels. • Sender driven The sender uses feedback reports from the receiver to learn network status and consequently adapt the outgoing bandwidth for that client. • Transcoder driven In transcoder based schemes, gateways are placed at strategic points in network to vary the quality based on the network status in each region. • Control data reduction and delivery schemes MCBA,a bandwidth smoothing algorithm is used Related Work

32. DART can deliver any form of time-sensitive metadata synchronously to the client. • DART has the advantage of being low on control overhead , network bandwidth sensitive and low startup overhead. • The proposed techniques in this paper can efficiently adapt to changes in the network to provide better QoS than schemes that consider sending rate synonymously with quality. Conclusion

33. Thank You!! 