[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [RRG] The use of UDP in LISP
Hi,
On 2007-12-13, at 20:49, ext Dino Farinacci wrote:
That pretty much leaves these options on how to go forward:
* use UDP with checksums
* use UDP-Lite with minimal checksum coverage
* push to update RFC2460, then use UDP with no checksums
* use a different protocol than UDP for tunneling
1, 2, and 4 are non-starters. The only option is 3 IMO.
I wouldn't want to eliminate 2 or 4 immediately.
IMO, 4 is an option, because UDP was explicitly designed as an end-to-
end protocol, i.e., not something that routers would need to implement
on the line card. If I correctly remember the discussions around
removing the IPv6 checksum and making transport checksums mandatory,
the desire to remove the IPv6 checksum was exactly to simplify router
processing, and shift the checksumming burden to the end hosts.
The issue we're running in now is that UDP's apparent attractiveness
as a router tunneling scheme conflicts with these earlier design
decisions. The reasons I have heard for why UDP is the right choice
for this are middlebox traversal and load-balancing. The former
shouldn't be an issue for LISP, because we're talking about router-
router tunnels. (AMT may be different.) The second can be achieved
through other means for another tunneling protocol.
Further, 2 is IMO an option, because I find it personally completely
acceptable that a major revamp of the routing architecture of the
Internet will require minor modifications to line cards. I know that
I'll get strong disagreement for this position, but I always felt that
the statement that LISP wouldn't require hardware modifications was
optimistic, and made at a time when the full consequences of the
proposal weren't understood
I have no preference among those four.
That's good. Now who is going to update the spec?
I provided text which Marshall has forwarded, who is going to insert
it? The IPv6 working group?
A short draft needs to be written that updates RFC2460, and it will
need to go through 6man and IESG approval. Your text is a good
starting point, but the real work will be the consensus building
required to make it go forward.
Lars
--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg