[e2e] Is a non-TCP solution dead?

Naidu, Venkata Venkata.Naidu at Marconi.com
Wed Apr 2 07:26:58 PST 2003


Hari,

-> My comments were restricted to dealing with wireless bit 
-> errors induced by  corruption.  Period.  I don't think having 
-> TCP deal with such link errors is a useful optimization.

  Did you have a chance to look at:
  TCP-DCR: A Novel Protocol for Tolerating Wireless Channel Errors.  
  "TCP-DCR delays responding to a packet loss indication by a small 
  period of time (one RTT) to allow channel errors to be recovered 
  by link level retransmission... An interested by-product of using 
  TCP-DCR is the inherent robustness it provides against ... 
  degradation due to packet re-ordering in the network." 
  http://dropzone.tamu.edu/techpubs/2003/TAMU-ECE-2003-01.pdf

Venkata.




More information about the end2end-interest mailing list