One potential issue is DOS attacks. Without loop detection, can I inject requests (adding Via's, say) that create additional work for proxies, by looping around until Max-Forwards kicks in? With a high Max-Forwards, I get a pretty good amplification factor.
Jonathan Rosenberg wrote: > > Jasson Casey wrote: > > > I've been going back through the RFC(s) and have a > > question. > > > > Requests can illegally loop or legitimately spiral. My > > current understanding is that an illegal loop is > > present when there is a corresponding via entry > > already in the via stack, and the message has a > > similar vector as with the previous via entry. > > > > What does Loop detection accomplish that cannot be > > accomplished with the Max-Forwards field? > > Indeed, a fine question, and the source of Open Issue #407, which > proposes to deprecate loop detection in favor of max-forwards. > > Loop detection has the benefit of detecting loops immediately, whereas > with Max-Forwards, they can loop around a few times till the counter > hits zero. However, I believe experience has shown us that the > complexities of detecting a loop vs. a legitimate spiral are > substantial, and the small imporvement in performance it provides is not > worth the cost. > > There has been little comment on the proposal to deprecate loop > detection; one email in favor, and a few questions. Without any > additional comment it will be removed from bis-06. > > Thanks, > Jonathan R. > > -- > Jonathan D. Rosenberg, Ph.D. 72 Eagle Rock Avenue > Chief Scientist First Floor > dynamicsoft East Hanover, NJ 07936 > [EMAIL PROTECTED] FAX: (973) 952-5050 > http://www.jdrosen.net PH: (973) 952-5000 > http://www.dynamicsoft.com > > _______________________________________________ > Sip-implementors mailing list > [EMAIL PROTECTED] > http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors _______________________________________________ Sip-implementors mailing list [EMAIL PROTECTED] http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
