Simon,
On 2010-04-16 05:40, Simon Perreault wrote:
> On 2010-04-15 11:22, Joel M. Halpern wrote:
>> However, a network can not give QoS treatment purely on the basis of
>> source and dest IPv6 address plus flow label. There simply is not enough
>> information. A client provided flow label is not a DSCP code point.
>
> My point with comparing this proposal with DSCP was that you cannot
> trust a host to do it in good faith in both cases. With DSCP, since
> hosts can trivially set the "fast bit", networks filter it at the edge.
> Same goes for the flow label if used for QoS.
If used as an *encoded* field for QoS. Exactly. That's why the idea
can be boiled down to:
For generic use, set a pseudo-random label per flow.
(forging it won't help you very much ;-)
For local use, set an *encoded* value meaning whatever you want,
but don't trust such values at the border.
I think we can simplify the proposal further after this discussion.
And then the decision is whether we want to allow the local use
case. I assure you there are people who want it.
Brian
>
> Other than that, I agree with your assessment.
>
> Thanks,
> Simon
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
