[e2e] Is a non-TCP solution dead?

Bob Braden braden at ISI.EDU
Fri Apr 25 12:41:09 PDT 2003


  *> 
  *> On Mon, 21 Apr 2003, Mark Allman wrote:
  *> 
  *> > Or, to flip this around.... when things are running smoothly, why
  *> > waste the network resources to send an ACK for every segment?
  *> 
  *> not just network resources, but cpu resources/state. If you're acking
  *> every segment received, you're using more resources immediately, but
  *> you don't need to run and expire a timer. It's a tradeoff that hasn't
  *> been fully explored.
  *> 
  *> 

Lloyd,

It was extensively explored in the early 1980s, and it was reconsidered when
RFC 1122 was written around 1988.


  *> 
  *> L.
  *> 
  *> that paper was my one attempt at being allman.
  *> frankly, allman's much better at it.
  *> 

;-))

Bob Braden




More information about the end2end-interest mailing list