OK - so this is within ISP networks. Could this be avoided by mapping
the DSCPs on entry and exit of their network? Do you know about CS1
within ISP networks? Or any impact at the edge?
Simon
On 5/24/2015 11:33 AM, Mikael Abrahamsson wrote:
On Sun, 24 May 2015, Simon Barber wrote:
My recent attention to DSCP has come from looking at what correct
mappings to 802.1D (now 802.1Q) would be. I have also run across a
couple of comments that legacy IP Precedence maps CS1 -> higher
priority than BE. Do you have any knowledge of how prevalent this
interpretation would be today, and whether it happens in any place
that would be a problem? (i.e. are there applications that would
generate these values, and rely on the behaivour, or routers that
mis-prioritize things at places that are likely a bottleneck)? I.E.
How important is it to consider these legacy behaivours today?
If ISPs today allowed DSCP marking to get propagated Internet wide and
didn't change their settings from what they have today, some would
treat AF1 and AF2 higher than BE, some would treat AF1 lower than BE
and AF2 higher than BE, some would treat AF1 and AF2 lower than BE.
That's why I'm saying AF1 and AF2 for less-than-BE isn't incrementally
deployable.
_______________________________________________
aqm mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/aqm
