SAMRG (Scalable Adaptive Multicast) IETF-78 Chairs: John buford and Thomas Schmidt Note-taker: Alexander Knauf 1. Introduction by Chairs 2. draft presentation - Application Layer Multicast Extensions to RELOAD (by Mario Kolberg) Thomas Schmidt: Question for Adoption was mentioned last meeting. Any comments? None. John Buford: Multicast out of scope from p2psip charter that’s why we adopt it here. Adopted as a RG document. 3. draft presentation - A RELOAD Usage for Distributed Conference Control (by Gabriel Hege) Joerg Ott: No Multicast mentioned in the presentation. Gabriel Hege: Charter of samrg contains group creation and management and this we do here for conferencing. Joerg Ott: Missing synchronization costs between focus peers, seem to be dominant. Gabriel Hege: Synchronization is done on conferencing event state package. Following proximity for topology creation optimizes synchronization cost. Joerg Ott: Missing state synchronization explanation. Gabriel Hege: The aim is to generate a global view on conference. Joerg Ott: Conference Event Package contains more than only membership information. I see problems while synchronization of conference event data in case of conflicts. Gabriel Hege: DisCo not a rich conferencing application, its P2P style. Joerg Ott: What does 'overloading' mean ? Usually, signaling does not overload a peer. Gabriel Hege: Overloading happens if a peer is in charge of media distribution and mixing, but this is still out of scope in the draft John Buford (as individual): Are focus peers allowed to be behind NATs? Gabriel Hege: Yes, since RELOAD is able for NAT traversal through ICE. John Buford (as individual): How large can a conference be? Gabriel Hege: Very large conferences probably do not scale because of media mixing, unless conversation is restricted. John Buford: Which overlay is used in RELOAD? What impacts are to the RELOAD base draft? Gabriel Hege: Our RELOAD usage is independent of the overlay algorithm. Our issue is with the access control policies. The standards do not fit our Usage. We need to define additional mechanism. 4. draft presentation - A common API for Transparent hybrid Multicast (by Matthias Waehlisch) John Buford: Did you get feedback on the draft? Matthias Waehlisch: Not yet, we are going to solicit reviews from now on. 5. Chairs presentation (John) - Update on the SAM testbed Thomas Schmidt: Will it be available for the SAM community? John Buford: Yes, we are working to make it available on the basis of research proposals. 6. draft presentation - Multiparty Transport Overlay Control Protocol (by Alexandru Petrescu) Joerg Ott: The previous work of UDP Multicast Tunneling Protocol (UMTP) seems to describe all techniques you have mentioned, it seems like you're tunneling? Alexandru Petrescu: We are not tunneling. So it’s not the same thing. Thomas Schmidt (as individual): Do you have made performance measurements in larger scale? Alexandru Petrescu: No, but its future work. In simulation or real. Thomas Schmidt (as individual): Is dynamic configuration of the tree planed? Alexandru Petrescu: I will pass this comment to the implementers. Lichun Li: How does a host join the group? Alexandru Petrescu: Via IGMP in multicast. MTO controller decides who receives what. 7. Chairs presentation - Document Status & Future Work Thomas Schmidt: We have one document adopted by the RG, another one in adoption call. What are the future plans for the MTOCP document? Alexandru Petrescu: We want to continue the project and work on the document. Thomas Schmidt: We are uncertain about future work on the XCAST documents, as no people are here. Thomas Schmidt: The RG plans to meet at Beijing and at Prag. John Buford: There will be an interim meeting at IEEE CCNC 2011 Las Vegas, Jan 7-11, 2011. Submission deadline is end of August. Thomas Schmidt; There is also a workshop "PerGroup" upcoming in IEEE Globecom'10, but it’s too late to submit papers.