new drafts of TCP-ENO, tcpcrypt

[Daniel B Giffin <[e-mail]>]

A few days ago we submitted new drafts of TCP-ENO and
tcpcrypt:

  https://datatracker.ietf.org/wg/tcpinc/documents/

The changes to TCP-ENO were trivial (adjust capitalization
of section titles and alphabetize acknowledgments).

The tcpcrypt document had quite a number of changes
(although most were cosmetic), addressing concerns from
these recent reviews:

  - SECDIR review (Stephen Kent)
  - GENART review (Dale Worley)
  - OPSDIR review (Zitao Wang)

The new draft also changes the mandatory-to-implement
key-agreement schemes from ECDHE with P-256 or P-521 to
*only* ECDHE with Curve25519.  It explains the rationale in
a new sub-section of "Security considerations".  But please
note that a final decision on what schemes to MTI has not
yet been made, so this may change.   WG members: Please see
the recent request for comments on the tcpinc list if you
would like to comment on this issue!

Apart from that, there were quite a lot of wording changes
and a few semantic tweaks.  Here's the diff:

  https://tools.ietf.org/rfcdiff?url1=draft-ietf-tcpinc-tcpcrypt-07&url2=draft-ietf-tcpinc-tcpcrypt-08

And lastly, here is a summary of the significant changes.

- Capitalize major words in section titles
- Always include an index with keys: mk[j], k_ab[j], k_ba[j]
- Cite the draft-ietf-tcpinc-api in Introduction and also
  where we require an interface to control session caching
- Cite RFC2104 for HMAC
- Note limit on output length of HKDF-Expand
- Try to make clearer that ENO negotiation chooses the TEP, and
  the TEP then determines the key-agreement scheme, KDF and
  associated parameters -- so all of these last could be varied
  by newly-specified TEPs
- At definition of k_ab, k_ba, insert explanation of which
  encryption key to use in a resumed session

daniel