[e2e] where end-to-end ends

Bob Braden braden at ISI.EDU
Tue May 22 18:12:03 PDT 2001

  *> I've always interpreted the ultimate "end" in 
  *> end-to-end to be the communicating processes 
  *> rather than the transport entities who are 
  *> providing services.


No, actually, I don't think that is the correct model for TCP.  There
was some consideration given to making TCP connections
process-to-process, but that was abandoned, I think because it seemed
to OS-dependent.  Instead, TCP connections end on "neutral territory"
between OS and network, the port space.  Each OS was then assumed to
have some way to couple its processes to ports.


More information about the end2end-interest mailing list