[apps-discuss] AD Fail - MPTCP API

[Pete Resnick <[e-mail]>]

If I were a good AD, I'd have been monitoring this all along and would 
have given a long heads up, before IETF Last Call. If I were a good AD, 
I would have asked for an Apps Directorate review. Now, Last Call 
complete and less than 19 hours before an IESG Telechat, I come to you, 
head hung in shame. Bad Pete.

https://datatracker.ietf.org/doc/draft-ietf-mptcp-api/ is on tomorrow's 
telechat. I love MPTCP and think it is exactly the right way to deal 
with multiple interfaces, but some of the stuff in this particular 
document makes me nervous. Unfortunately, I have not written any code to 
a sockets API in too long. I have already bugged two people for comments 
(thanks Chris and Cyrus), but I wouldn't mind more people looking at 
this document. It's only Informational, but I am quite sure people will 
start coding this stuff. I'm personally more worried about the legacy 
API interactions (see section 4.2.2 in particular), but if you have 
thoughts on the new APIs as well, they would be welcome.

I'm going to be writing up my comments, but at this point I'm not 
planning on asking to DISCUSS it. But if anyone has concerns where you 
think, "Holy crap! You can't let people start coding that, even if it is 
just experimental!", you've got not much time to let me know.

Begging your forgiveness,

pr

-- 
Pete Resnick<http://www.qualcomm.com/~presnick/>
Qualcomm Technologies, Inc. - [phone]