At 08:31 PM 9/15/2011, Joel M. Halpern wrote:
I am the assigned Gen-ART reviewer for this draft. For background on
Gen-ART, please see the FAQ at
<http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq>.
Please resolve these comments along with any other Last Call
comments you may receive.
Document: draft-ietf-tsvwg-sctpsocket-31.txt
Sockets API Extensions for Stream Control Transmission Protocol (SCTP)
Reviewer: Joel M. Halpern
Review Date: 15-Sept-2011
IETF LC End Date: 27-Sept-2011
IESG Telechat date: 6-Oct-2011
Summary: This document is ready for publication as an Informational RFC
Major issues: None
Minor issues:
For the shepherd's consideration: Is it sufficiently clear how
the source address will be selected when sending a message, if
bindx or wildcarding has been used to bind multiple source
addresses? My guess is that this lack of description matches other
specs in how that is handled, but I wanted to check.
Here's consecutive responses from the two main authors/editors:
"
> Yes, the lack of description matches what you find in other
specifications at least
> specs to my knowledge. I think this "fuzziness" in the spec is so
that implementors
> have room to play.. which they do ;-)
The only thing we can add is a statement that the source address will be
* one of the bounded addresses, if the socket is bound to
non-wildcard addresses
* one of the addresses of the host, it the socket is bound to a
wildcard address.
Maybe add some text regarding IPv4/IPv6.
However:
* I think all of this is obvious.
* The crucial point, how to select a specific source address, does not
belong to an API spec. Maybe in the base protocol spec. But it was
left out (I think on purpose), since we do not require the SCTP
to do source address selection. It might be done by the IP layer.
For example, if you use a userland implementation.
"
What do you think Joel?
James
Nits/editorial comments:
Editorial gs/proceeded/preceded/
_______________________________________________
Gen-art mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/gen-art
