Hi, Brian,
On 23/1/21 18:47, Brian E Carpenter wrote:
[...]
At the end of the day, I guess we cannot publish a PS that clearly
breaks E2E, while at the same time claim or pretend that we keep/have
E2E....
Much as I had concerns about draft-ietf-spring-srv6-network-programming,
I think we should be clear about what e2e is about. It *isn't* about
idealised transparency or even about forbidding packet mangling.
For example, what we said in RFC1958 is:
"The basic argument is that, as a first principle, certain required end-
to-end functions can only be performed correctly by the end-systems
themselves."
That wasn't the last word, of course: see RFC3724, for example.
Well, header removal is one of those. At least at the time of this
writing, it breaks, at least:
* ICMPv6-based error reporting
* IPsec
[...]
available at the final destination. But IMHO it *doesn't* tell us
that a routing header must not be deleted by the penultimate
hop, because it's of no value to the final destination anyway.
(Logically, it's also a no-op, because the final destination
ignores routing headers. Go figure.)
Which gives even more the impression that we ended up screwing-up the
architecture because folks want to use RH, and their own boxes fail to
comply with even the most basic requirements for RH handling...
Regards,
--
Fernando Gont
SI6 Networks
e-mail: fgont@xxxxxxxxxxxxxxx
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492