[anonsec] 3401 and highjacking
hartmans-ietf at mit.edu
Mon Mar 13 11:08:15 PST 2006
>>>>> "Stephen" == Stephen Kent <kent at bbn.com> writes:
Stephen> What SRTP did was to add integrity and confidentially to
Stephen> RTP in a very careful, space-efficient fashion. In so
Stephen> doing, they reduced overhead well below what one could
Stephen> achieve via use of DTLS or ESP or AH. This is not a
Stephen> practice I encourage for application in general, because
Stephen> the costs of custom designs of this sort are very high,
Stephen> and for most apps the bandwidth savings will not justify
Stephen> these costs. Also, as Joe noted, the likelihood of
Stephen> introducing security problems in such designs (and in
Stephen> implementations) is also very high.
The multimedia application also has a number of other special
requirements. I think that multimedia streams over cellular networks
are one of the few cases I've seen where confidentiality without
integrity may be reasonable.
Humans will know if an attacker significantly corrupts a voice or
video stream. However the cost of lost packets do to corruption of a
MAC may be unacceptable given radio error rates.
More information about the ANONSEC