David A. Bryan P2PSIP AdHoc Meeting IETF 64 (Vancouver)

draft-bryan-sipping-p2p-01 Did not iterate draft for this IETF Another iteration soon Moving forward, I expect this will change a great deal to adapt and change to reflect protocol decisions made in this meeting and future (?) meetings

New Work Evaluating routing techniques –Looking at a limited set of P2P approaches –Which common approaches are specifically appropriate for real time communications? –Simulations of performance in communications situations of many P2P algorithms –Much of this work is academic, but the non- academic part of this work is likely to result in an informational draft for input to decision making on finding appropriate routing technique

Routing Techniques Preliminary results –We can eliminate a large number of existing P2P approaches designed for file sharing Don’t guarantee termination Optimized for number of hits, not lookup speed Have never actually been implemented (far more common than you might imagine) –Because we have buddy lists, may be able to take advantage of “social networking” in overlay approach Can be significant performance improvement (see Sprout algorithm)

Routing Techniques Need to ensure system supports redundancy –In some approaches designed for file sharing, this hasn’t been as much of a concern since multiple copies are assumed We aren’t likely to find a perfect or “best” technique – evaluating each and every one is likely to be a rat hole –New techniques every day In Summary –The criterion we need for communications is likely to be quite different than for file sharing

More New Work Use Cases Draft(s?) –Have an extensive use cases draft After discussion at this IETF have several more to add plan to combine this work with Eunsoo Shim Will release soon –Possibly one draft on immediate uses and another of more speculative/long term uses Speculative might be best presented in P2PRG