The Use of Timed Efficient Stream Loss-Tolerant Authentication (TESLA) in the Secure Real-time Transport Protocol (SRTP)
draft-ietf-msec-srtp-tesla-05

[Ballot discuss]
There is a normative downreference from this document to the tesla
document that was not noted in the last call.  If we're going to have the downref policy at all, this is the situation where it should be used: a new specification making a downreference to an informational specification that is critical to its understanding and that has not received as much review for clarity as a standard.

Sorry to delay things, but their does seem to be community support behind our current policy.

[Ballot comment]
I gave the equivalent of a Discuss (before the document was deferred from 9/29 telechat):
1. the 32-bit SRTP MAC needed explanation, as to why such a short tag, quite
    forgeable, was an acceptable risk (I sent this before the telechat as private mail
    to Baugher, with background of SRTP weak MACs)
2. the document should be given a review by the AVT WG.  I sent a message to the
    WG mailing list and received several replies supporting the document.
Mark B has previewed an 05 draft which addressed 1. and I'll put in a noob when
that draft appears.

[Ballot comment]
I gave the equivalent of a Discuss (before the document was withdrawn? deferred?):
1. the 32-bit SRTP MAC needed explanation, as to why such a short tag, fairly
    forgeable, was an acceptable risk (I sent this before the telechat as private mail
    to Baugher, with background of SRTP weak MACs)
2. the document should be given a review by the AVT WG.  I sent a message to the
    WG mailing list and received several replies supporting the document.
Mark B has previewed an 05 draft which addressed 1. and I'll put in a noob when
that draft appears.

At the request of the authors, this document is being removed from the 2005-09-29 telechat agenda.  They believe that the GenART Last Call comments deserve another look, and a revised I-D will probably be forthcoming.