Multimedia Multicast on the Internet

Abderrahim Benslimane is Professor of Computer Science and Engineering at the University of Avignon, France. Currently, he is responsible for the Masters programme in networks, telecoms and multimedia. He is team leader of the computer networks and multimedia applications research group. His research and teaching interests are group communication protocols, quality of service in wired and mobile networks and inter-vehicular communication. He is also author of several refereed publications in these areas, as well as being involved in related scientific projects.

• Preface xviiChapter 1. Multicast Routing on the Internet 1Jean-Jacques PANSIOT1.1. Introduction and definitions 11.2. Multicast addressing 41.2.1. Limited scope addressing 51.2.2. GLOP global addressing 51.2.3. Dynamic addressing: MALLOC 61.3. Structure of a multicast router 71.3.1. The unicast routing base for multicasting (MRIB) 71.3.2. Tree information base (TIB) 81.3.3. Multicast forwarding information base (MFIB) 81.4. Relationship with the other protocol layers 101.4.1. Relationship with the lower layer 101.4.2. Relationship with the upper layers 121.5. Belonging to groups: IGMP 121.5.1. IGMP version 1 131.5.2. IGMP version 2 131.5.3. IGMP version 3 141.6. Routing in flood-and-prune mode and the RPF 151.6.1. Reverse path forwarding or RPF check 151.6.2. Pruning 161.6.3. Protocol cost 171.6.4. DVMRP 171.6.5. Mbone 181.6.6. PIM dense mode: PIM-DM 181.7. Link-state routing and MOSPF 181.7.1. MOSPF principle 181.7.2. MOSPF inter-areas 191.7.3. Cost of MOSPF 201.8. Routing with explicit construction: PIM-SM and CBT 201.8.1. PIM sparse-mode principles: PIM-SM 211.8.2. Discovery of RPs: boot strap routers (BSR) 241.8.3. Maintenance of the PIM-SM tree 241.8.4. Core based trees: CBT 251.8.5. Bidirectional PIM 251.8.6. Cost of explicit methods 261.9. Inter-domain multicast routing 271.9.1. MASC/BGMP architecture 271.9.2. BGP multiprotocol extensions 281.9.3. Interaction with intra-domain routing 291.9.4. BGMP 291.9.5. PIM-SM and MSDP solution 301.10. Model of multicasting with a single source: SSM 321.10.1. Express 321.10.2. The SSM and PIM-SM model 331.10.3. Limitations of PIM-SSM 331.11. Multicasting and IPv6 341.11.1. IPv6 multicast addressing 341.11.2. Protocol for group subscription: MLD 351.11.3. RP-embedded mechanism 351.12. Other multicast routing proposals 361.12.1. Simple multicast 371.12.2. Logical addressing and routing: LAR 371.12.3. Reunite 381.12.4. Hop by hop multicast routing: HBH 391.13. Comparison of various protocols 401.13.1. Quality of the broadcast trees 401.13.2. Cost of protocols 421.14. Alternatives to multicast routing 431.14.1. Multiple unicast connections 431.14.2. Multicasting for small groups 431.14.3. Application level multicast 431.15. Conclusion 441.16. Bibliography 441.17. Glossary of acronyms 49Chapter 2. Hierarchical Multicast Protocols with Quality of Service 51Abderrahim BENSLIMANE and Omar MOUSSAOUI2.1. Introduction 512.2. Multicast principle 532.2.1. Advantage of multicasting 532.2.2. Technological constraints 552.2.3. Main types of trees 562.2.3.1. Shared tree/specific tree 562.2.3.2. Shortest path tree (SPT) 572.2.3.3. Steiner tree 572.2.3.4. Centered tree (CBT) 582.2.3.5. Summary 582.3. Multicast routing protocols 592.3.1. DVMRP 592.3.2. PIM 602.3.3. MOSPF 612.3.4. IP multicast 622.3.5. Limitations of the current multicast routing protocols 632.3.5.1. DVMRP 632.3.5.2. PIM 632.4. Quality of service in multicast routing 642.4.1. SJP 642.4.2. QoSMIC 662.4.3. QMRP 672.4.4. Conclusion 682.5. Hierarchical multicasting 692.5.1. HDVMRP 702.5.2. LGC 732.5.3. HIP 742.5.4. QHMRP 782.5.5. Conclusion 812.6. Hierarchical structure for multicasting 822.6.1. Context of the system 822.6.2. Construction of local groups 822.6.2.1. Construction of the neighborhood 822.6.2.2. Construction of transit groups 832.6.2.3. Grouping and election 832.6.3. Construction of hierarchical trees between servers 842.6.3.1. Use of centered trees 852.6.3.2. Use of SPT trees 872.6.3.3. Comparison between the two methods 882.6.4. Management of the hierarchical structure 892.7. Conclusion 902.8. Bibliography 90Chapter 3. A Transport Protocol for Multimedia Multicast with Differentiated Quality of Service 93David GARDUNO, Ernesto EXPOSITO and Michel DIAZ3.1. Introduction 933.1.1. Multimedia 933.1.2. Partial QoS 933.1.3. Multicast 953.1.4. Text organization 963.2. State of the art 963.2.1. Point-to-point multimedia data transmission 963.2.1.1. UDP and TCP 963.2.1.2. SCTP 973.2.1.3. DCCP 983.2.1.4. Networking layer: IntServ 983.2.1.5. Networking layer: DiffServ 993.2.2. Multicast algorithms 1003.3. Network model, Tree and QoS oriented multicast service 1023.3.1. Introduction 1023.3.2. Hierarchized graph 1043.3.3. Degree Bounded Shortest Path Tree (DGBSPT) 1073.3.4. Model and simulations 1163.4. Fully Programmable Transport Protocol 1183.4.1. Introduction 1183.4.2. Design principles 1193.4.3. Contextual model of QoS 1193.4.3.1. QoS specification 1193.4.3.2. QoS mechanisms 1203.4.4. Protocol specification 1213.4.5. Implementation and evaluation 1233.5. Integration of multicast services and multimedia protocols 1253.5.1. Deployment of transport services by proxies 1253.5.1.1. Basic FPTP architecture and mechanisms 1263.5.2. The M-FPTP multimedia multicast service 1283.5.3. Tests and results 1303.6. Conclusion 1313.7. Bibliography 132Chapter 4. Reliability in Group Communications: An Introduction 135Vincent ROCA4.1. Introduction 1354.2. Which reliability for which applications? 1364.2.1. Reliability levels 1364.2.2. Group models 1374.2.3. Transmission models 1374.2.4. Multiplicity of applications and their needs 1384.3. Challenges and big classes of solutions in the case of a reliable group communication service 1394.3.1. Challenges 1394.3.2. Reliable scaling and communications: problems 1404.3.3. Scaling of control traffic 1404.3.3.1. Use of removal mechanisms by recipients 1404.3.3.2. Use of FEC codes 1414.3.3.3. Use of assistance node trees 1424.3.4. Scaling of retransmissions 1424.3.4.1. Use of FEC 1424.3.4.2. Use of a retransmission server tree 1424.3.4.3. Local retransmissions 1424.3.5. Considering the heterogenity 1434.3.6. First assessment 1444.4. FEC codes 1444.4.1. Codes for packet erasure channels 1444.4.2. The concepts of systematic codes and MDS codes 1454.4.3. Classification of FEC codes 1454.4.4. Small block codes 1464.4.4.1. Principles 1464.4.4.2. Problem linked to block segmentation 1464.4.4.3. Use in the reliable communication systems 1474.4.5. Large block codes 1474.4.5.1. Introduction 1474.4.5.2. Operation mode of LDPC-staircase and LDPC-triangle codes 1474.4.6. Rateless codes (also known as extensible codes) 1524.4.6.1. Introduction 1524.4.6.2. Principles of online codes 1524.4.6.3. Comparison with the LDPC-staircase and triangle codes 1534.4.7. A few additional notes on the FEC rateless and large block codes 1534.5. Conclusion 1544.6. Bibliography 155Chapter 5. End-to-end Approaches for Reliable Communications 157Vincent ROCA5.1. Introduction 1575.2. The main protocol classes and the block approach of the IETF 1585.3. The FEC building block 1595.3.1. The “FEC encoding ID” and “FEC instance ID” 1595.3.2. The FPI (FEC payload ID) 1595.3.3. The “FEC object transmission information” (FEC OTI) 1605.3.3.1. Block partitioning algorithm 1615.3.3.2. The n algorithm 1625.4. The NORM approach 1635.4.1. Operating principles 1635.4.1.1. General ideas 1635.4.1.2. Main types of packets 1635.4.1.3. Transmission window mechanism 1645.4.2. The building blocks used 1655.4.2.1. FEC block 1655.4.3. Scope 1665.5. ALC approach 1665.5.1. Operating principles 1665.5.1.1. General ideas 1665.5.1.2. Close-up on the layered transmission principle 1675.5.1.3. And if we used only one layer? 1695.5.2. The building blocks used 1695.5.2.1. The LCT block 1705.5.3. Scope 1715.6. The FLUTE file transfer application on ALC 1725.6.1. Operating principles 1735.6.2. An example of FDT instance 1745.6.3. Scope 1755.7. A few NORM and FLUTE/ALC available implementations 1765.8. Conclusion 1775.9. Bibliography 177Chapter 6. Router-assist Based Reliable Multicast 181Prométhée SPATHIS and Kim THAI6.1. Introduction 1816.2. Motivations and objectives 1836.3. Protocol network architecture 1866.3.1. Active error recovery (AER) and light-weight multicast services (LMS) 1866.3.2. Pragmatic general multicast (PGM) 1876.3.3. Active reliable multicast (ARM) and multicast actiffiable (MAF) 1876.4. Classification 1886.4.1. Organizing the control tree 1886.4.2. Repair entities 1906.4.3. Local approaches 1936.4.3.1. Receiver-initiated approach 1936.4.3.2. Sender-initiated approach 1946.4.4. Buffer management 1956.4.4.1. Receiver-initiated approach 1956.4.4.2. Aggregated ACKs 1966.4.5. Exposure of receivers 1976.4.5.1. ARM and PGM 1976.4.5.2. MAF 1996.4.5.3. AER and LMS 1996.4.6. Feedback implosion 2026.4.6.1. Aggregation 2026.4.6.2. Optimization of aggregation 2036.4.7. Suppression 2056.4.7.1. Anticipation 2056.4.7.2. LMS and MAF 2056.4.8. Loss recovery burden 2066.4.8.1. ARM and PGM 2066.4.8.2. AER and LMS 2076.4.9. Standardization of router-assist based approaches 2086.5. Placement mechanisms 2096.5.1. Motivations and objectives of the placement of repair entities 2106.5.2. Location models 2116.5.3. Applications of the p-median problems to the placement of repair entities 2126.6. Performance analysis 2136.6.1. Large scale simulations and experiments 2136.6.2. Analytical models 2146.6.3. Precursory works 2156.6.4. Comparative analytical studies of router support approaches 2156.7. Conclusion 2166.8. Bibliography 217Chapter 7. Congestion Control in Multicast Communications 223CongDuc PHAM and Moufida MAIMOUR-BOUYOUCEF7.1. Introduction 2237.2. Congestion control 2257.2.1. Congestion control: a bit of theory 2257.2.2. The congestion control in practice: example with TCP and the AIMD process 2267.3. The congestion control in group communications 2297.3.1. Information filtering and representativeness 2297.3.2. Scalability 2317.3.3. Heterogenity management 2327.3.4. In brief 2337.4. Single-rate approaches 2337.5. Multi-rate approaches 2357.6. Approaches with router assistance 2397.7. Conclusion 2427.8. Bibliography 2427.9. Appendix 1: summary table of the approaches quoted in this chapter 2457.10. Appendix 2: acronyms of the protocols presented 246Chapter 8. Approaches to Multicast Traffic Engineering 247Christian JACQUENET8.1. Introduction 2478.2. The use of DiffServ mechanisms 2498.2.1. Reminder of the DiffServ architecture 2498.2.2. Risks of over-use of resources within the DiffServ domain 2508.2.3. Marking and signaling: establishment and maintenance of multicast distribution trees with differentiated qualities of service 2508.3. Multicast traffic engineering and MPLS networks 2578.3.1. The difficulty of activating multicast traffic processing capabilities in MPLS domains 2578.3.2. Multicast traffic engineering using the point-to-point LSP MPLS resources 2588.3.2.1. Establishment of multicast distribution trees at the edge of MPLS networks 2588.3.2.2. Construction of distribution trees according to the service classes supported in the MPLS domain 2618.3.3 Multicast traffic engineering using point-to-multipoint LSP MPLS tree structures 2628.3.3.1. Establishment of point-to-multipoint LSP 2628.3.3.2. Routing of multicast flows in traffic-engineered point-to-multipoint LSP trees 2678.4. Conclusion 2688.5. Bibliography 269Chapter 9. Towards New Protocols for Small Multicast Groups: Explicit Routing and Recursive Unicast 271Ali BOUDANI and Abderrahim BENSLIMANE9.1. Introduction 2719.2. Explicit multicast routing protocols 2739.2.1. Xcast 2739.2.2. Xcast+ 2759.2.3. Advantages and disadvantages of the Xcast technique 2769.2.3.1. Advantages of the Xcast technique 2779.2.3.2. Disadvantages of the Xcast technique 2779.2.4. Generalization of the Xcast technique 2799.2.4.1. Description of the GXcast protocol 2799.2.4.2. Links between GXcast and the maximum transfer unit 2819.2.5. Incremental deployment of an Xcast protocol in a network 2819.2.5.1. Tunneling 2819.2.5.2. Premature X2U 2839.2.5.3. Semi-permeable tunneling (only with IPv6) 2839.2.6. Different explicit multicast propositions 2849.2.6.1. SGM 2859.2.6.2. CLM 2859.2.6.3. MDO6 2869.2.6.4. Somecast 2869.2.6.5. ERM 2869.2.6.6. MSC 2869.2.6.7. DCM 2879.2.7. Summary and limitations of the various explicit multicast routing protocols 2879.3. Recursive unicast 2909.3.1. REUNITE 2929.3.2. HBH 2939.3.3. SEM 2959.3.4. Comparison between HBH and SEM 2979.3.5. SREM 3009.4. Conclusion 3049.5. Bibliography 304Chapter 10. Secure Multicast Communications 307Melek ÖNEN, Refik MOLVA and Alain PANNETRAT10.1. Introduction to multicast security 30710.1.1. Multicast applications and their characteristics 30710.1.2. Security requirements 30910.1.3. Limitations of the unicast solutions 31010.2. Multicast authentication 31110.2.1. Definition and requirements 31110.2.2. Techniques using symmetric algorithms 31210.2.2.1. Multicast message authentication codes (MMAC) 31210.2.2.2. TESLA 31310.2.3. Combination of asymmetric and symmetric algorithms 31510.2.3.1. Hash trees 31510.2.3.2. Hash chains 31610.2.3.3. The use of erasure codes 31810.2.4. Conclusion 32010.3. Multicast confidentiality 32010.3.1. Definition and requirements 32010.3.2. Re-encryption trees 32210.3.2.1. Iolus 32210.3.2.2. Cipher sequences 32410.3.3. LKH: Logical Key Hierarchy 32610.3.4. Conclusion 32710.4. Reliability of key distribution protocols 32810.4.1. Requirements 32810.4.2. Solutions based on replication techniques 32910.4.3. Solutions based on the use of FEC 33010.4.4. Conclusion 33010.5. General conclusion 33110.6. Bibliography 332Chapter 11. Scalable Virtual Environments 335Walid DABBOUS and Thierry TURLETTI11.1. Introduction 33511.2. Specificities of the LSVE 33711.2.1. Scalability 33711.2.2. Interactivity 33811.2.3. Heterogenity 33811.2.4. Consistency 33911.2.5. Reliability 33911.3. Multipoint limitations 34011.3.1. Routing 34011.3.2. Subscriptions and unsubscriptions latency 34111.4. SCORE-ASM 34211.4.1. Assessment of the additional cost related to the use of multipoint 34311.4.2. The role of the agents 34411.4.2.1. Association of multipoint cells-groups 34611.4.2.2. Assignment of multipoint groups 34611.4.3. Communications in SCORE-ASM 34711.4.3.1. Communication between participants 34811.4.3.2. Participants-agent communication 34911.4.3.3. Communication between agents 35011.4.4. Connection to the virtual world 35111.4.5. Subscriptions update mechanism 35111.4.6. Clipping algorithm 35211.4.7. Conclusions regarding SCORE-ASM 35311.5. SCORE-SSM 35411.5.1. Problematic 35511.5.2. Choice of design 35611.5.3. SCORE-SSM structure 35611.5.3.1. Filtering 35711.5.3.2. Heterogenity and multimedia flow 35811.5.3.3. Correspondence with the network multipoint 35911.5.4. Prospects regarding SCORE-SSM 35911.6. Final comment 36011.7. Bibliography 361List of Authors 363Index 365