[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lemonade] QUICKSTART update

Tony Finch wrote on 4/19/07 19:56 +0100: 
draft-fanf-smtp-quickstart-a-00 and draft-fanf-smtp-quickstart-b-00 have
been published. I've separated it into two documents to make the two
different profiles clearer. I prefer the -B profile but I'd like opinions
from others about the trade-off between complexity and effectiveness.


Taking off my area director hat, and speaking as a technical participant:

It's my opinion that the way profile B interacts with the TLS layer is problematic for implementations. The server code I work on uses the NSS (Mozilla) SSL/TLS library. While it would be possible to implement profile B using that SSL/TLS library, it's very difficult as it requires writing a special NSPR I/O layer that slots below the SSL stack and deals with the STARTTLS server response after the client has activated the TLS state machine. Requiring tricky code like that for security integration is never a good idea.

The problem is that activating the TLS layer is a software state change that takes over both socket directions. So it really shouldn't happen until the only subsequent protocol will be TLS packets. Having secure and non-secure packets in-transit at the same time is not a good idea from a state management point of view.

If you want additional input on the topic, I suggest asking on the TLS WG mailing list.

               - Chris



_______________________________________________
lemonade mailing list
lemonade at ietf.org
https://www1.ietf.org/mailman/listinfo/lemonade
Supplemental Web Site:
http://www.standardstrack.com/ietf/lemonade