Jarno,
I completely agree!
Intserv would work fine with ANY other mechanism for selecting flow
label values. It does not have to be PRN.
Intserv seems to work fine with FILTER_SPEC C-type 1, and 2, i.e. based
on "port numbers", which are not PRN.
Furthermore, work on classification mechanisms has made such a progress,
that "hashing", as lookup for flow state information, as suggested in
Appendix A, RFC 2460, is obsolete.
Even more, IMHO, it is not the business of a IETF standard, to tell what
mechanism to implement for classification, whether is "hashing" or
something else.
Regards,
Alex
[EMAIL PROTECTED] wrote:
>
> Brian wrote:
> > The distinction made in the proposal by the MSB is between a
> > pseudo-random
> > value as defined in RFC 2460 Appendix A (for intserv) and a
> > non-random value
> > (for diffserv). But they are both e2e values, unlike the
> > DSCP. This should
> > be documented.
> >
>
> But who need any of the values to be pseudo-random? Intserv would work just
> fine if the flow label values would be taken out of the sequence 1,2,3,4,...
> Intserv signaling should provide for no problem of re-use after reboot
> either, since after a reboot the signaling would be done again before any
> use of the flow label.
>
> Jarno
S/MIME Cryptographic Signature
