Multicast: Difference between revisions
reorganize ip multicast |
what is web conferencing? define it first |
||
Line 8: | Line 8: | ||
* [[Internet Relay Chat]] |
* [[Internet Relay Chat]] |
||
* [[PSYC]] |
* [[PSYC]] |
||
* [[Web conferencing]] |
|||
* [http://wwcn.org WWCP] |
* [http://wwcn.org WWCP] |
||
Revision as of 09:02, 19 September 2005
- Multicast is sometimes used to refer to a multiplexed broadcast
Multicast is the delivery of information to a group of destinations simultaneously using the most efficient strategy to deliver the messages over each link of the network only once and only create copies when the links to the destinations split. By comparison with multicast, conventional point-to-single-point delivery is called unicast. When unicast is used to deliver to several recipients, a copy of the data is sent from the sender to each recipient, resulting in inefficient and badly scalable duplication at the sending side. Many popular protocols like XMPP accept this shortcoming however, and limit the amount of possible recipients instead.
Multicast technologies on the Internet
- IP Multicast (described in this document)
- Internet Relay Chat
- PSYC
- WWCP
The word "Multicast" is typically used to refer to IP Multicast, which is a protocol for efficiently sending to multiple receivers at the same time on TCP/IP networks, by use of a multicast address. It is also commonly associated with audio/video protocols such as RTP.
But there are also other protocols on the Internet that implement the multicast concept. Some connection-oriented link layers, such as Asynchronous Transfer Mode, have built-in mechanisms for "point-to-multipoint" or "multipoint-to-multipoint" connections. This model generally assumes that the stations participating in the communication are known ahead of time, so that distribution trees can be generated and resources allocated by network elements in advance of actual data traffic. By contrast, the IP Multicast model has been described by Internet architect Dave Clark as "You put packets in at one end, and the network conspires to deliver them to anyone who asks."
While IP Multicast is an extremely satisfying conceptual model, it requires a great deal more state inside the network than IP unicast model of best-effort delivery does, and this has been the cause of some criticism. What is worse, no mechanism has yet been demonstrated that would allow the IP Multicast model to scale to millions of senders and millions of multicast groups as would be necessary to make fully-general multicast applications practical in the commercial Internet. As of 2003, most efforts at scaling multicast up to large networks have concentrated on the simpler case of single-source multicast, which seems to be more computationally tractable.
For both this reason and also reasons of economics, IP Multicast is not in general use in the commercial Internet. Other multicast technologies, which are not based on IP Multicast, are quite popular however, like Internet Relay Chat and PSYC. They may not be as elegant, but they are pragmatic and scale better for large amounts of small groups. See below for a list.
IP Multicast
However, some communities within the public Internet make regular use of IP Multicast (see the Mbone for an example), and IP Multicast is used for special applications within private IP networks such as Internet2.
Link local multicast, where IP Multicast packets are sent to groups of hosts on the same physical or virtual data link layer, does not require complex routing, and is therefore much more widely deployed. It is used in IPv6 for address resolution, and in zeroconf networks for service discovery, name resolution, and address conflict resolution, replacing inefficient broadcast protocols.
IP multicast conferencing was first demonstrated on a wide scale when it was used to broadcast several sessions from the 23rd IETF in March, 1992 to researchers and interested observers around the world. Selected IETF sessions have continued to be multicast over the MBONE and private multicast networks since then.
Multicast security is a major issue. Standard, practical, communications security solutions normally employ symmetric cryptography. But applying that to IP Multicast traffic would enable any of the receivers to pose as the sender. This is clearly unacceptable. The IETF MSEC workgroup is developing security protocols to solve this problem, mostly within the architectural framework of the IPsec protocol suite.
IPsec cannot be used in the multicast scenario because IPsec Security associations are bound to two hosts and not many. IETF proposed a new protocol TESLA, which is quite convincing and flexible for multicast security.
Concerning IP Multicast, there are three basic kinds of delivery:
Modern multicast algorithms are often capable of combining and managing the first two.
IP Multicast protocols:
- Internet Group Management Protocol (IGMP)
- Protocol Independent Multicast (PIM)
- Distance Vector Multicast Routing Protocol (DVMRP)
- Multicast OSPF (MOSPF)
- Multicast BGP (MBGP)
- Multicast Source Discovery Protocol (MSDP)
- Multicast Listener Discovery (MLD)
See also
- Spanning tree (networks)
- Flooding algorithm
- Anycast
- NBMA Nonbroadcast Multiple Access Network
- VideoLAN - a free software multicasted video streaming application
- Session Announcement Protocol
External links
- IPMSI IP Multicast Standards Initiative
- RFC 3170
- Multicast over TCP/IP HOWTO
- Nemo - Resilient Overlay Multicast
- FatNemo - Emulating FatTrees for Overlay Multisource/Multicast
- End System Multicast
- BBC Multicast trial stream list (currently only works with some UK based ISPs who are taking part in the trial, also lists unicast streams for BBC radio and News 24)