Cross posting to MBONED.
Thanks for elaborating. Yes, there is some extra work. But
assuming LISP and other UDP tunnels will limit the UDP-Lite
checksum coverage to the minimum (i.e., just the header), things
become very lightweight. (But yes, not as lightweight as not doing
any checksumming.)
Light weight but not light weight enough. Still a non-starter for
product development.
As long as LISP is a research effort, it is very free in following
the published RFCs. But as soon as we're starting to talk about
standardization, it is important that LISP follow the RFCs that
define the behavior of the protocols that it uses. (After all,
you'd want people to follow the MUST and SHOULDs in any future
LISP RFCs, right?)
And we have to make standards practical for vendors to implement or
vendors sway from the standard and create defacto standards. Please
don't take this lightly. This is really important to make protocols
implementable and not too expensive where non-standard alternatives
are sought.
The pressing point, as Marshall points out, is AMT. This issue is
holding AMT from going forward on standards track. We need a
resolution soon.
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 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?