Jarno,
[EMAIL PROTECTED] wrote:
>
> Alex,
>
> I don't think it has been shown yet, that the diffserv architecture actually
> needs any specific support from the IPv6 flow label.
The needs of the architecture is materialized through the needs of the
mechanisms. This has been shown repeatedly, and I thought you were here.
> While the
> MF-classification is mentioned in the diffserv architecture, the value of
> that on the administrative boundaries is less than clear.
The benefit of the MF classifiers is very clear. The draft has mentioned
also the benefit for access networks. I thought you've read the draft.
> An SLA between two domains could well be specified to only consider DSCP
> values, and the associated PHB, volume, charging, etc.
>
> Could someone please explain why this could not work, and what specifically
> is the value of signaling the PHB in the flow label field?
>
> Jarno
>
I am afraid you're starting a vicious circle. Please reread Brian's
messages. It is also significant that using a IANA number was discussed,
and proposed - the recent draft is a refining of an earlier proposal.
Alex
S/MIME Cryptographic Signature
