1 / 5

to TCP or not to TCP

to TCP or not to TCP. D. Towsley U. Massachusetts. throughput. loss prob. Observations on TCP. TCP  1/p ½ (defined as a response function) not appropriate for region of low p  redefine response function for small p ~ 1/p 0.84 p < 0.001 (S. Floyd)

Download Presentation

to TCP or not to TCP

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. to TCP or not to TCP D. Towsley U. Massachusetts

  2. throughput loss prob. Observations on TCP • TCP  1/p½(defined as a response function) • not appropriate for region of low p  redefine response function for small p ~ 1/p 0.84 p < 0.001 (S. Floyd) or ~ 1/p , p < 0.001 (T. Kelly) • both do well! • at least 1Gbps • fairness? Enhanced TCP CC algorithm robust

  3. the hard stuff! Observations on Transport • several successful transport protocols • even TCP (carefully tuned)! • > 1 GBs • current limitations lie outside transport protocol • access BW (1GE) • NIC issues • I/O (disk to NIC) • OS issues

  4. HSTCP? throughput loss prob. Observations on Transport (ctd.) • need for standardized protocol • reduces net admin costs  frees up money for research • specialized vs. general purpose • scalable protocols • 10Gbs – 40 GBs – 160Gbs – 640Gbs • bit errors? • congestion notification? • startup, fairness 1/p

  5. What do workloads look like?

More Related