Le 15 avr. 2010 à 00:26, Brian E Carpenter a écrit :
> Or we can strongly recommend that all hosts set the flow label, so
> that we can use the 3-tuple {source address, dest address, flow label}.
>
> What do people think?
In view of Mark's remark about the impossibility for routers to identify
5-tuples of all IPv6 packets, I tend to think this is the right way to go.
But then another requirement of RFC 3697 should be relaxed:
- The RFC says: "To avoid accidental Flow Label value reuse, the source node
SHOULD select new Flow Label values in a well-defined sequence"
- This implies a stateful assignment of flow labels.
A stateless alternative, where flow label values are just a hash of the 5-tuple
(or 3-tuple if alone to be available), should IMHO be permitted, and not less
recommended than the stateful well-defined sequence:
- It permits to compute flow labels *on a per datagram basis*, just before
their being transformed into packets.
- Hash functions may be different in different hosts (advices may be welcome,
but there is no need to standardize)
Regards,
RD
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
