PhD Seminar • Systems and Networking — Conflux — Multi-homed Adaptive Bitrate Protocol for On-site, Live Video Streaming (WIP)Export this event to calendar

Friday, July 26, 2019 1:30 PM EDT

Sharon Choy, PhD candidate
David R. Cheriton School of Computer Science

Television broadcasters use live streaming to capture developing and live events, which are then streamed to the broadcaster’s respective television station for post-processing before being delivered to viewers. In this scenario, a high video bitrate is needed for broadcasting purposes, low latency is required for interactive events such as interviews, and video stalls must be avoided as they are detrimental to the viewer’s experience. 

One innovation that industry has employed is the use of multi-homing. The availability of additional links can be used to improve the quality and reliability of the video stream. However, using multi-homing has many challenges including knowing when to trade-off video quality for reliability and determining when it is worthwhile to incur the extra bandwidth cost to use multiple links. Furthermore, because video is typically streamed over LTE/wireless in this scenario, it is especially challenging as the condition of the available links may be highly variable. 

In this talk, I present my progress on Conflux, which is a multi-homed adaptive bitrate protocol for on-site, live video streaming. Conflux models the multi-homed streaming problem as a decision making under uncertainty problem. Conflux makes decision regarding the overall video bitrate and determines how to partition data over multiple links. We outline the challenges that we have faced and the work we plan to complete in the future. 

We welcome any feedback and suggestions regarding our system.

Location 
DC - William G. Davis Computer Research Centre
1304
200 University Avenue West

Waterloo, ON N2L 3G1
Canada

S M T W T F S
29
30
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
1
2
  1. 2024 (96)
    1. April (19)
    2. March (27)
    3. February (25)
    4. January (25)
  2. 2023 (296)
    1. December (20)
    2. November (28)
    3. October (15)
    4. September (25)
    5. August (30)
    6. July (30)
    7. June (22)
    8. May (23)
    9. April (32)
    10. March (31)
    11. February (18)
    12. January (22)
  3. 2022 (245)
  4. 2021 (210)
  5. 2020 (217)
  6. 2019 (255)
  7. 2018 (217)
  8. 2017 (36)
  9. 2016 (21)
  10. 2015 (36)
  11. 2014 (33)
  12. 2013 (23)
  13. 2012 (4)
  14. 2011 (1)
  15. 2010 (1)
  16. 2009 (1)
  17. 2008 (1)