Brian E Carpenter wrote:
True, but while we're being pragmatic, it seems clear to me that the
RRG
will never reach consensus on a host based solution in any case. And
since we have one in progress in the IETF already, aren't we done
here?
Brian,
no doubt that discussions in RRG won't change how widely Shim6 will be
deployed. But this does not mean that RRG should ignore the concerns
that people have raised against Shim6. If RRG came up with a host-based
solution that mitigates these concerns, then this would be a big
achievement in my opinion.
And RRG can already claim some success in this regard: One concern that
has been raised against Shim6 is the lack of network control on traffic
engineering. In RRG, at least two new host-based solutions have been
proposed that mitigate this concern: Six/One gives the network explicit
traffic engineering control through address prefix rewrites in routers.
And multi-path TCP gives the network an implicit means for traffic
engineering through bandwidth limits and packet loss.
Another disadvantage of Shim6 -- which applies just as well to other
host-based solutions that provide a "stable address" -- is the
complexity of the extra indirection layer: This, plus the cryptographic
techniques by which the new indirection is secured, are not trivial.
They demand special expertise from network administrators, and they
limit network design freedom because they require certain address
configuration methods. As I have tried to explain in Minneapolis, a
hostname-oriented stack architecture would mitigate these issues -- by
replacing the "stable address" with a human-readable hostname that maps
onto regular, non-cryptographic IP addresses. The stack architecture
would thus become simpler, and more friendly for users, administrators,
operators, and application developers.
So overall, while Shim6 is for sure a great technology, there is room
for improvement. And here, I believe, is where RRG can contribute.
- Christian
_______________________________________________
rrg mailing list
[email protected]
https://www.irtf.org/mailman/listinfo/rrg
