[tsvwg] Comment on draft-ietf-tsvwg-transport-encrypt-13

[Tom Herbert <[e-mail]>]

Regarding new text on Ipv6 extension headers, the draft states:

"o On the one hand, protocols do not necessarily have an incentive to
expose the actual information that is used by the protocol itself and
could therefore manipulate the exposed transport header information to
gain an advantage from the network. The incentive to reflect actual
transport header information has to be considered when proposing a
method."

I think this is a mostly irrelevant argument because:

- The QUIC spin bit, which is touted as the best example of a non-TCP
protocol exposing transport layer information, is not used by the end
protocol hence hosts can set it to try to gain an advantage as the
argument suggests. It's even worse than that, since the spin bit is in
UDP payload it's quite possible that the network misinterprets a
packet that is not QUIC (but uses the same port number), in which case
such packets may be at a disadvantage.
- STT (https://tools.ietf.org/html/draft-davie-stt-08) is an example
of how the TCP protocol itself can be spoofed explicitly for the
purpose of fooling the network into provide TCP-level service for
something not at all TCP (in this case IP tunnels).
- The argument precludes the possibility that the transport layer
information in Hop-by-Hop options could be authenticated by the
network and hence a contract is established between hosts and network
that the information is honest and correct. FAST
(https://tools.ietf.org/html/draft-herbert-fast-04) gives a
description of that.

Fundamentally, transport layer is end-to-end information. There is no
contract between end hosts and the network that hosts have to be
honest or correct in setting information in the transport layer-- the
only contract is between the endpoints. There is incentive for hosts
to set transport information per the networks implicit requirements,
but that's not because we are interested in getting any of the
benefits listed in the draft, it's simply because if we don't abide by
the networks implicit requirements our packets get will likely get
dropped somewhere in the path. Ultimately, this is what has driven
host stacks and applications to the least common denominator, and is
exactly why any new transport protocols will encrypt as much as
possible.

Tom