On 02/02/2017 09:26 PM, Brian E Carpenter wrote:
On 02/02/2017 22:14, Fernando Gont wrote:
...
The current impossibility to parse an IPv6 header chain that includes
unknown Next Header values
Wait... you're talking about parsing by intermediate systems. The destination
node either can understand the new Next Header or transport protocol,
or it can't. That's fine, and is the intended result.
If it can, it already knows the syntax, so what's the point of a uniform
syntax? -- it could be anything, and wouldn't change anything.
results in concrete implications for the
extensibility of the IPv6 protocol, and the deployability of new
transport protocols. Namely,
o New IPv6 extension headers cannot be incrementally deployed.
o New transport protocols cannot be incrementally deployed.
In both cases, add "in the presence of interfering middleboxes".
This is not new. For a document moving from PS to Standard, it is
not something we can change.
Note, I am all for coming back to this problem, after we have the
Internet Standard in place. Maybe we can fix it or maybe we can't,
but it's IMHO off topic here.
Ok. Makes sense.
Cheers,
--
Fernando Gont
SI6 Networks
e-mail: fgont(_at_)si6networks(_dot_)com
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492