AW: [Geopriv] Location Conveyance: Scenarios
"Tschofenig, Hannes" <[email protected]>
| Newsgroups | gmane.ietf.sip,gmane.ietf.geopriv |
|---|---|
| Message-ID | <A5D2BD54850CCA4AA3B93227205D8A30898A81@MCHP7IEA.ww002.siemens.net> |
Hi Henning, > Thanks, Hannes; these are helpful to focus the discussion. > > I think, however, that the issue of security (encryption or > integrity) > and conveyance are essentially orthogonal. In other words, > you can apply > the same techniques, with similar problems, in all three conveyance > mechanisms (by-value, by-attachment, by-reference). I fully agree with you. I don't recall that the document ever discussed the possibility of protecting a location-by-reference. Hence, I was puzzled why the encryption is a big deal in one case and not an issue at all in the second one. > The procedural > details differ (such as XML-DSIG and XML encryption vs. > S/MIME vs. TLS), > but the strength and the certificate problems are similar, albeit not > identical. (For example, it is probably somewhat harder to deal with > multi-recipient cases in S/MIME than in other mechanisms.) I agree with you. > > It might help streamline the discussion if we can agree on > this general > statement or show that it is mistaken. > Ciao Hannes > Henning > > Tschofenig, Hannes wrote: > > Hi all, > > > > I have compiled a few message flows that would make it > easier to discuss > > the different scenarios. I have listed some challenges > along with each > > of the scenarios. > > > > > > 1) Location by End Host / Per Value/ in Body > > > > UA Alice SIP Proxy UAS-A UAS-B UAS-C > > > > | SIP Message | | | | > > |---------------------------->| | | | > > | Location (value/body) | | > > | |----------------->| > > | | | > > | SIP Message | > > |<-----------------------------------------------| > > | | > > > > Problems: > > * Likelyhood of the UA using S/MIME; nice on paper but does > it reflect > > reality? > > * UA Alice needs to know that (a) intermediaries do not > need to process > > the PIDF-LO > > (b) recipient has a certificate > > * Unclear what happens if there are multiple recipients. > > > > > > > > 2) Location by End Host / Per Reference / in Body > > > > > > UA Alice SIP Proxy UAS-A UAS-B UAS-C > > > > | SIP Message | | | | > > |---------------------------->| | | | > > | Location (ref./body) | | > > | |----------------->| > > | | | > > | Dereference | > > |<==============================================>| > > | | > > | | > > | SIP Message | > > |<-----------------------------------------------| > > | | > > > > Problems: > > * Can be protected using S/MIME (see issue above) > > * If not protected then the considerations of who is allowed to > > dereference to a PIDF-LO are applicable. There is certainly > a difference > > between a reference pointing to a PIDF-LO at the UA itself and a > > reference obtained using one of the discussed Geopriv L7 protocols. > > It might be difficult for the end host to make the PIDF-LO > available. It > > might be difficult for him to have a server-side certificate but it > > might be easier to control access to it. > > > > > > 3) Location by SIP Proxy / Per Reference / in Header > > > > UA Alice SIP Proxy UAS-A UAS-B UAS-C > > > > | SIP Message | | | | > > |---------------------------->| | | | > > | | SIP Message | > > | |----------------->| > > | | Location | > > | | (ref./header) | > > | | | > > | | | > > | | Dereference | > > | |<================>| > > | | | > > | SIP Message | > > |<-----------------------------------------------| > > | | > > | | > > | | > > > > Problems: > > > > * What should the SIP proxy but into the PIDF-LO when it is created > > (policies, identifiers, etc.)? > > * Authorization aspects: see above (Why cannot just > everyone that sees > > the reference in the header passing by dereference it?) > > * Signing of the reference theoretically possible. SIP > Proxy, however, > > needs to know the intended recipient. > > > > > > 4) Location by SIP Proxy / Per Value / in Header > > > > UA Alice SIP Proxy UAS-A UAS-B UAS-C > > > > | SIP Message | | | | > > |---------------------------->| | | | > > | | SIP Message | > > | |----------------->| > > | | Location | > > | | (value/header) | > > | | | > > | SIP Message | > > |<-----------------------------------------------| > > | | > > | | > > | | > > > > Problems: > > > > * What should the SIP proxy but into the PIDF-LO when it is created > > (policies, identifiers, etc.)? > > * Data URI approach possible but with increased message size. > > * Signing of the PIDF-LO possible but approach similar. SIP Proxy, > > however, needs to know the intended recipient. > > > > > > 5) Location by SBC / Per Value / in Body > > > > UA Alice SIP Proxy UAS-A UAS-B UAS-C > > (SBC) > > | SIP Message | | | | > > |---------------------------->| | | | > > | Location (value/body) | SIP Message | > > | |----------------->| > > | | Location | > > | | (value/body) | > > | | | > > | SIP Message | > > |<-----------------------------------------------| > > | | > > | | > > | | > > > > Problems: > > > > * If there is no solution for the cases available where the > SIP proxy > > can add a PIDF-LO and a reference is not acceptable (because of the > > latency) it is likely that boxes will just do it. > > * Does not play "nicely" with SIP. > > * Problems with E2E security > > > > > > 6) Location by End Host / Per Referene / in Header > > > > UA Alice SIP Proxy UAS-A UAS-B UAS-C > > > > | SIP Message | | | | > > |---------------------------->| | | | > > | Location (ref./header) | | > > | |----------------->| > > | | | > > | Dereference | > > |<==============================================>| > > | | > > | | > > | SIP Message | > > |<-----------------------------------------------| > > | | > > > > Problems: > > > > * Not sure about S/MIME usage > > * Simlar considerations as in (2) > > > > Finally, I think it would be useful to differentiate > between location > > based services in general and emergency services since the > requirements > > are, in most cases, different. > > > > Ciao > > Hannes > > > > _______________________________________________ > > Geopriv mailing list > > [email protected] > > https://www1.ietf.org/mailman/listinfo/geopriv > _______________________________________________ Sip mailing list https://www1.ietf.org/mailman/listinfo/sip This list is for NEW development of the core SIP Protocol Use [email protected] for questions on current sip Use [email protected] for new developments on the application of sip