1 / 6

Congestion Window Validation

IETF-88, Vancouver, 2013. Congestion Window Validation. d raft-ietf-tcpm-newcwv-03 G. Fairhurst, A . Sathiaseelan , R . Secchi School of Engineering University of Aberdeen, Scotland, UK. Updates since draft -02. Response to congestion in non-validated phase Start of recovery:

yepa
Download Presentation

Congestion Window Validation

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. IETF-88, Vancouver, 2013 Congestion Window Validation draft-ietf-tcpm-newcwv-03 G. Fairhurst, A. Sathiaseelan, R. Secchi School of Engineering University of Aberdeen, Scotland, UK

  2. Updates since draft -02 • Response to congestion in non-validated phase • Start of recovery: cwnd = (Max(pipeACK,LossFlightSize))/2 • End of recovery: cwnd = (Max(pipeACK,LossFlightSize)-R)/2 ‘R’ is the retransmission volume • pipeACK considered for ‘End of Burst losses’ • Validated phase when pipeACK is undefined • at the initial stage and after recovery

  3. ExptLinux newCWV module • Module follows the latest draft • Implements the example pipeACK calculation • Linux kernel versions > 3.0 required • Code file and compilation instructions:https://github.com/rsecchi/newcwv • Feedback very welcome

  4. newCWV with different applications HTTP streaming (DASH player) 3 persistent connections to server Video streaming (VLC) FTP application (Filezilla) – cwnd reduction of one connection after NVP of 5 mins

  5. Issues to address … • Non Validated Period (NVP) • We must decay after a “long” time • Path changes, apps going idle, etc. • Is 5 mins too long to halve cwnd? Or OK? • 0.19% HTTP connections in one hour (CAIDA traces) • Do we need pacing of line-bursts in NVP? • Max burst method does not work with no ACK-Clock

  6. Summary • We think this is NEARLY ready!!! • We need people to experiment with it… • We need some help to finish… • Finally, thanks to: • This work was partly funded by the EC under its 7th Framework Programme through the RITE project (ICT-317700) • Zia Hossain helped with analysis and Linux Coding • Thanks to all who commented on-list and off-list!

More Related