260 likes | 354 Views
SIP Musings. Henning Schulzrinne SIP Summit – Las Vegas May 8, 2002. Outline. Where is SIP in 2002? challenges: multimedia conferencing binding SIP to other protocols privacy and identity emergency services SIP tenets challenged by the real ($) world. Where in the world is SIP?.
E N D
SIP Musings Henning Schulzrinne SIP Summit – Las Vegas May 8, 2002
Outline • Where is SIP in 2002? • challenges: • multimedia conferencing • binding SIP to other protocols • privacy and identity • emergency services • SIP tenets challenged by the real ($) world
Where in the world is SIP? • SIP as the signaling protocol for future applications • 3GPP • Cable modems (DOCSIS DCS) • IM: AOL interworking (eventually...), Windows Messenger • but: H.323 dominates videoconferencing, trunk replacement • Proprietary protocols dominate for Ethernet phones • Slow uptake of VoIP
How did we get there? • Not quite what we had in mind • initially, for initiating multicast conferencing • in progress since 1992 • still small niche • even the IAB and IESG meet by POTS conference… • then VoIP • written-off equipment (circuit-switched) vs. new equipment (VoIP) • bandwidth is (mostly) not the problem • “can’t get new services if other end is POTS’’ “why use VoIP if I can’t get new services”
How did we get here? • VoIP: avoiding the installed base issue • cable modems – lifeline service • 3GPP – vaporware? • Finally, IM/presence and events • probably, first major application • offers real advantage: interoperable IM • also, new service
SIP developments in 2001/02 • SIP revision (“RFC2543bis”, RFC3261) almost done: • semantically-oriented rewrite • layers: message, transport, transaction, transaction user • SDP extracted into separate draft • UA and proxy have the same state machinery • better Route/Record-Route spec for loose routing • no more Basic authentication • few optional headers (In-Reply-To, Call-Info, Alert-Info, …) • Integration of reliable provisional responses and server features • DNS SRV modifications
SIP developments in 2001/02 • SIP revision backwards compatible • “new” messages work with RFC 2543 implementations • some odd allowed RFC 2543 behavior no longer allowed • CPL finished – merger with iCal • sip-cgi published • IM & presence mostly done, except for IM sessions (over TCP) separate message stream
SIP developments in 2001/02 • Work continues on staples: • early media (announcements) • resource reservation (COMET) • SIP security • SIP events • User identification: "network asserted identity", privacy • Call transfer and call control • Now three SIP working groups: • SIP for protocol definition and extensions • SIPPING for applications and “vetting” • SIMPLE for IM & presence
Conferencing • SIP designed for (multicast) multimedia conferences • most conferences are small (~5) central-server ("star") conferences • may be star of stars • single-source multicast (SSM) • need supporting infrastructure: • conference management • user management • floor management • user management
Conferencing • Conference management • create, delete • properties (visibility, duration, ...) • User management • mass invitation • automated and manual user admission • join and leave notification • membership lists • Floor control • request and release floor • re-order queue
Binding SIP to other protocols • SIP not a universal data transport protocol connect SIP to other protocols via URLs • get more data • Call-Info, Error-Info, Alert-Info • NOTIFY: pointer to message content • obtain configuration information • signed (long) version of SDP • publish data • presence document • buddy lists • CPL, sip-cgi, servlet • conference (user, floor) management
Binding SIP to other protocols • Sometimes needed for address-of-record • REGISTER addresses AOR • sometimes for particular UA only • OPTIONS addresses endpoint(s) • clients need to obtain URIs & set them • service mobility
Identity and privacy • Conflicting goals: • caller may want privacy • callee wants real identity SIP as telemarketer's dream? • FBI wants CALEA • Traditional "caller-ID" doesn't work well in SIP: • user, not device identity • iffy notion of trusted providers • SIP direct no need for carrier • Proposal in SIP* WG: • asserted identity without authentication • signed identity based on first-hop authentication
Emergency services • Two types of emergency services: • emergency calls ("911", "112") • emergency notification ("inverse 911") • emergency calls are hard: • PSTN gateway dialing 911 may be located anywhere, far away from IP phone • VPNs IP address may not reveal network location
Emergency services • Work for both "legacy" PSAPs and IP-based PSAPs • find location of phone based on • street address geocoding long/lat • longitude, latitude • find appropriate PSAP for jurisdiction • find ESR (routing number) call appears like 911 call • convey location
Emergency notification • notify public officials and citizens of emergencies: • "tornado coming" • "fugitive alert" • current systems are typically single-mode (fax, telex, phone, TV, loudspeaker) • don't scale well • very limited information content • don't reach citizens outside calling area • people at work • hard to authenticate
SIP-based emergency notification • SIP has scalable event notification feature • use for hierarchical notification reflecting civil lines of authority • use XML/WSDL message bodies to semantically describe emergency: • location • type of emergency • instructions • ... • allow automated reaction: • routing to legacy systems (pagers, police radios) • translation
Domesticating SIP • Temptation to make it look like SIP on the wire, but lacks core SIP design properties • SIP is designed as a common carrier protocol: • content-neutral • application-neutral • plausible deniability in proxies • proxy as minimally aware entities: • pass unknown SIP headers • handle any method • don't touch bodies
SIP design principles • SIP design: • extensibility in all elements methods, headers, content types • disclosure of message handling to user Via • user influence over message handling Require, caller preferences
SIP: mind-changing • "network" services end-system services • proxies as willing helpers, not service disablers • trusted network or network federations PKI, distributed trust, "tunneled" trust • media-neutrality
3G: Falling into the WAP trap? • "3G adopts SIP" "3G adapts SIP" • if just another IP transport, should not need to explicitly do anything • wireless-specific: bandwidth scarcity compression • overspecified no need for C/S/P-CSCF as standard just one of many possible architectures • allow roofed + walled garden with customed characters and "forever wild" national park
The fallacy of topology hiding • "don't reveal server structure" • infinite ways to leak to a determined "spy" • proxies don't reveal IP-layer connectivity or bandwidths • knowing proxy addresses doesn't reveal proxy capacity • security through obscurity? • costs of hiding: • lack of diagnosability of faults • extensibility decreases
SIP doesn’t have to be in a phone sip:lava@cs.columbia.edu
Conclusion • Goal of single world-wide, transport-independent session setup protocol within reach • Danger of non-interoperable islands • repeat of mobile signaling division • ISDN • Need to address operational issues • not signaling load, administrator load • emergency calls • trust & privacy