Folks, while attempting to use draft-williams-on-channel-binding in
the SASL working group, we came across an ambiguity.
In response to IETF last call comments we added the concept of a
unique prefix and a registry of prefixes for channel binding type. We
added a requirement that applications make sure that one channel could
not conflict with another channel type. However we didn't specify how
the prefix was to be used.
This ambiguity made using specifications more complex than needed.
So, we propose to actually say that the prefix needs to be a prefix.
This change has the support of the authors, myself, and members of the
SASL community including the author of the document trying to use this
mechanism.
In particular, we propose adding the following text:
>> "Under this framework, channel bindings MUST start with the
>> channel binding unique prefix followed by a colon (ASCII 0x3A).
>> "
The document is currently in auth48. I will approve this change if
there are not objections in a week.
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf