>>Only if we define a standard way of compressing the ports and protocol type
>>into the flow label.
>       i don't think there needs to be a standard way.  if there's some
>       pseudorandom and unique number in flow label field (as suggested in
>       RFC2460) it is more than enough.

        oops, what I meant here was:
        - router/RSVP device looks at full 5-tuple, on the first packet it sees,
          and remember flow label value
                (need to chase extension headers)
        - for subsequent packets router/RSVP device will classify packets
          based on flow label value
        there's no need for compression rule from 5-tuple to flow label
        (some other protocol may use something other than 5-tuple!)

        I admit that ESP case is tricky.

itojun
--------------------------------------------------------------------
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