Kit Plummer wrote:
On May 3, 2005, at 3:36 AM, Brian E Carpenter wrote:
Kit Plummer wrote:
On Apr 24, 2005, at 8:55 AM, Ran Liebermann wrote:
Ofcourse if the label will also be used to classify traffic for
diffserv then this is one more reason for SPs to override the label to
zero on all ingress traffic from customers.
If this is the case, and I hope it is...wouldn't is simply make
more sense for the SP to ignore the FL all together? Therefore, if
the FL were pertinent on the other end...
The diffserv use case would use an e2e flow label to allow all diffserv
classifiers along the path to correctly reclassify the packet, even if
the protocol and port numbers are obscured by encryption. That only
works
if the flow label value is preserved right up to the last router on the
last LAN on the path.
Brian
Brian,
Just trying to see this straight...and I think I am understanding
correctly. But, my question is why would a SP override the label, if
the label is to provide information from a classification perspective?
The answer as I see is that they shouldn't...and as you pointed out the
necessity for preservation from true e2e.
The point is that (at the ISPs' request) the diffserv model allows
each ISP along the path to apply a different policy - so a packet
marked for EF treatment inside one ISP might be marked for AF
treatment inside another ISP. (You can argue that would be stupid, but
that's what the diffserv WG heard from the ISPs.) So different SPs will
interpret the same label differently.
Brian
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
