> >   > > I had a preliminary idea and was going to write something
  > >   > > about it. Basically the sender can generate the flow label
  > >   > > based on a hash of the source and destination port numbers.
  > >   > > This is straight forward enough. If we want to make it
  > >   > > more sophisticated then we can add another number
  > >   > > to the hash input (e.g P1 || P2 || x).
  > >   > > Where x can be something specific to this flow. But
  > >   > > I think P1 and P2 are enough.
  > >   >
  > >   > I guess this doesn't work where the source is setting up
  > >   > flows to more than
  > >   > one destination using the same source and destination port
  > >   > numbers. So I
  > >   > think you need to include destination address in the hash
  > >   > computation as
  > >   > well.
  > > 
  > > => Well, there is no requirement that the flow label is
  > > globally unique. So this should work as long as you use
  > > a unique value between 2 addresses.
  > > 
  > 
  > However, it's only useful for a subset of cases, 

=> I didn't really understand what you mean by 'subset
of cases'.


so I don't think
  > it affects the basic definition in the draft.
  > 

=> Not at all, it's not meant to affect the basic
definition in the draft, on the contrary it's
meant to complement it. I actually like the 
draft's proposal.

Hesham
  >    Brian
  > 
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------

Reply via email to