Kenjiro Cho wrote:
>
> Jun-ichiro itojun Hagino wrote:
> > So, kjc wanted to at least identify each of the flow, at least for
> > statistics purposes, <...>
> > on (b).
>
> minor correction:
>
> My point was that leaving the flowlabel field unused doesn't do any
> good for QoS/IPv6 deployment.
I agree.
> So, I proposed to itojun to set flowlabel by default because I think
> the availability of flowlabel values will promote practical use of
> this field.
>
If Intserv-only is "good" for QoS/IPv6, the use with both Intserv, and
Diffserv, is
"more than good", is an enhancement, a plus.
> The original definition may not be perfect but at least there are
> useful applications. It isn't only for intserv; e.g., it can be used
> for best-effort traffic to provide better fairness by WFQ-style
> schedulers.
Adding the ability to do Diffserv QoS would improve the applicability,
the useful applications,
while allowing all of the original applications. It is a win-win
situation.
> -Kenjiro
Alex
S/MIME Cryptographic Signature
