1 / 8

Multi-rate Congestion Control

Multi-rate Congestion Control. Tapan Karwa CS590F. Multi-rate Data Organization. Simulcast : Different multicast groups, each sending all the data/signal at different rates. Cumulative : Layered division of data. Each layer on a different mcast group. Higher layers refine data. Related Work.

janice
Download Presentation

Multi-rate Congestion Control

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Multi-rate Congestion Control Tapan Karwa CS590F

  2. Multi-rate Data Organization • Simulcast : Different multicast groups, each sending all the data/signal at different rates. • Cumulative : Layered division of data. Each layer on a different mcast group. Higher layers refine data.

  3. Related Work • RLM : • Uses a state machine. Adaptive timer controlled experiments used to decide subscription. • Shared learning. • Digital Fountain Approach : • Combine Tornado codes with layered mcast, to enable ‘digital fountain’. • Start drinking at any time and adapt drinking rate to available BW. Need any k out of (k+l)

  4. Related Work • RLC : Receiver-driven Layered CC • Congestion control is achieved using SP’s. Burst of data sent right before SP. • Receivers make join attempts only after a SP. Allows new users to quickly ramp up by placing SP’s conveniently. • BW doubles with every higher layer and number of SP’s decrease with every higher layer. • Synchronizes receivers.

  5. Technique • All these schemes wait for loss before they react. • Use ECN mechanism : • Don’t have to wait for loss to detect congestion. • No wastage of resources already used by packets. • Significance for streaming video/audio. • No loss computation at the receiver. Hence, faster. • Current routers are normally in the regime of probabilistic marking. Effect of using ECN mechanisms. • Not looked at yet!

  6. Problems to be Dealt With • How many marked packets should the receiver wait for – One or more? • TCP suggests considering a marked packet as a dropped packet and reacting accordingly. Other mcast CC schemes also wait for a single packet loss to react. • Solutions : • Wait for a single marked packet. Too aggressive in dropping especially considering probabilistic marking. • Wait for a subset proportional to data rate of highest layer. What subset? ½? Effect on TCP flows? Is this an engineering problem?

  7. Things to be Measured • Quality as indicated by Loss Rates : • Effect on TCP flows. • Convergence Time • Stability in terms of layers subscribed to.

  8. Future Avenues • Sender-based schemes using ECN information to dynamically alter sending rates. • Using Diffserv mechanisms with ECN. • Other proxy based schemes.

More Related