D079 Supporting Internet Multicast Multimedia. UCL Report for 1st Quarter 1997. Jon Crowcroft & Peter Kirstein 0. Abstract In this report, we cover the last quarter's work in the project on multicasting multimedia. The report is in 9 subsections: [1]DARPA Multicast Multimedia Work - Introduction [2]Multicast Multimedia - Specific Goals [3]Overall Status [4]Multicast Routing: Background [5]QoS - RSVP and Aggregation [6]Reliable Multicast [7]Reliable Multicast Protocol [8]Key Issues [9]Next Steps 1. DARPA Multicast Multimedia Work 1.0 Goals The goals of the project are (as per the proposal): 1. Mechanisms for supporting multicast routing protocols such as core placement strategies for CBT and PIM, Quality of Service path selection in multicast tree formation, and unicast support for reverse path calculation. 2. Mechanisms for utilising multicast capability in applications such as feedback control in a multicast environment, management of applications in a multicast group, and reliable multicast transport. 3. Mechanisms for integrating resource management with multicast forwarding, such as interface between forwarding engine and caching issues. The purpose of the proposed research project is to develop novel mechanisms for supporting and utilising Internet multicast. 1.1 First 6 Months - In the first year, we have devised a hierarchical version of PIM to address the RP/Core placement problem, This has been fed via IDMR WG of the IETF into the design of baseline PIM, and into CBT (c.f. Ordered CBT). - We have devised a specific reliable multicast protocol for use with distribute Multicast transport protocol from the LBL wb too, but with extensions based on accurate clock synchronisation to provide for lower latency (but higher network load) repair. To be presented at the joint UCB/EPFL workshop on Multimedia Networking at Lausanne in July.[J. Crowcroft: Distributed Virtual Reality Multicast Protocol - We have completed the Session Directory, and Invitation protocols, together with releases of implementations. [M. Handley: Session Invitation Protoco, ; M. Handley: Session Directory Announcement Protocol, Internet Drafts] - We have Documented the overall Architecture in a paper to be presented at INET 96 [Crowcroft and Handley: The Internet Multimedia Conferencing Architecture] 1.2 Second 6 Months - OCBT is work at UCSC which takes ideas from HPIM, and applies them to CBT, removing looping problems, and possibly improving scaling and core placement. Reported in IDMR at Memphis - YAM - this is work by Ken Carlberg from SAIC, PhD student at UCL published in ACM CCR. B-MART is a bulk transfer multicast protocol, being proposed for pre-loading web mirror sites, and VRML world databases. - Security for SDP/SIP, Rat - Windows ports of Rat, Nte, Vic 2. Multicast Multimedia * Multicast Enhancements for: + Mobility + QoS (Tree Selection) + Reliability + RSVP/Int-Serv Flow State Aggregation + Scaling of Group Management Traffic 3. Overall Status * Project Fully Staffed Zheng Wang -> Bell Labs Panos Gevros, Labros Lambrinos ?? + Progress on Track + Liason with CAIRN coming on stream + Collaboration with ISI/Berkeley/MIT in place + RSVP and RED deployed in UK (UCL+BT) and moving to intl links + IPv6 beta from Cisco. 4. Multicast Routing: Background * Tree Selection (Ken Carlberg - paper in ACM CCR) for QoS (Delay or Thrughput) * Mobility - Taxonomy (Zheng Wang, Philippe Oechslin, Mark Handley): * Solution for Mobile Source Based Tree - router based state: keep tracking history as source address changes inherit RPF+Prune state 5. QoS - RSVP and Aggregation * Partial Congruent Tree and Flow Aggregation - idea for merging router state - + trades off against partial delivery of multicast data to destination leaf networks that have no receivers - + advantages in stat muxing of aggregated non correlated flows on shared, but distribution trees for disjoint flows! measurement based signaling and admission 4. Reliable Multicast * Atanu Ghosh, Lorenzo Vicisano * Defined Reliable Multicast Framing Protocol + Analogous to RTP, but for non real-time data + SRM and RMP and BMART profiles + Proposed in IRTF new RM Working Group + Prototyped in Java + in the berkeley simulator, ns 5. Reliable Multicast Protocol * BMART - + uses general packet FEC + has TCP like (in sense of adaption loop) congestion avoidance, but receiver based for scaling + Supports heterogeneity (similar to redundant and layered audio coding RTP profiles) + Prototype in C++ and in Java + tested in wide area 6. Key Issues * Scaling Experiments on + RSVP Proxy, and RSVP aggregators (NAT like implementation planned to re-write group and unicast addresses so that state can be CIDRized) + RM experiments with BMART to do WWW mirroring, and maybe DIS contribution 7. Next Steps * Bring up CAIRN link * Measure BMART congestion avoidance * Get some RSVP large scale experience * Look at RTP and RMFP over Controlled Load, and over RED (already some measurements of both of these) * Write up for SIGCOMM 98 or INFOCOM. Appendix Deliverable Status 1. A core placement algorithm suitable for CBT and PIM, and a technical report detailing the operation of the algorithm. HPIM is part of this 2. A technical report on QoS support in multicast routing. YAM is part of this 3. A technical report on the secuirty issues and protection mechanisms for multicast networks. TBD - partly in Perkins' work on secure SDR etc. 4. A algorithm for unicast support for reverse path calculation, and a report detailing its operation. TBD 5. A technical report on the algorithms and performance data of various feedback control algorithms. TBD 6. A technical report on the lessons learnt from design of reliable multicast transport protocols. C.f. papers at HPCS and HIPPARCH by L Vicisano, Z.Wang, Crowcroft et al 7. Software for management and control video conferencing system and a report detailing its design and implementation. TBD 8. Software for kernel implementation of integrated resource management and multicast forwarding. Measurement RSVP ideas address this, but need development further. and subsequent implementation.