I have reviewed this draft. It is well written, justifies its
recommendations and I believe it is ready to move forward. I have asked
for an IETF Last Call. However, I did see one editorial issue and one
technical issue that I would like you to discuss and correct if
necessary (even during the Last Call):
If there is a high
proportion of tunnel traffic, traffic will not be distributed as
intended across the paths between R1 and R2.
... high proportion of traffic from one or small number of tunnels,
traffic will not be ...
* Intermediate IPv6 router(s) will presumably encounter a mixture
of tunneled traffic and normal IPv6 traffic. Because of this,
the design may also include {protocol, dest port, source port}
as input keys to the ECMP and/or LAG hash algorithms, to
provide additional entropy for flows whose flow label is set to
zero, including non-tunneled traffic flows. Whether this is
appropriate depends on the expected traffic mix and on
considerations of implementation efficiency.
Really, I don't think we are constructing routers that could only route
tunnel packets. A device such as a router needs to be capable of
handling both tunnel and non-tunnel packets. Indeed, the router will in
general not even be able to know if it is processing tunnel packets. I
would recommend changing the above "may also include" to "should also
include", and striking the last sentence.
Jari
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
