Core Issue
IP multicasting provides a method to send information to a group of hosts at the same time, which is different from the normal IP routing that sends information only to a specific host (unicast) or to all hosts (broadcast) at one time. IP multicasting conserves bandwidth by sending the packets transmitted by a source to a group of receivers simultaneously, instead of separate unicast transmissions or broadcasting traffic.
Resolution
The Internet Assigned Numbers Authority (IANA) has assigned the Internet Protocol (IPv4) Class D address space in the range from 224.0.0.0 through 239.255.255.255 to be used for IP multicast. Some of the addresses in this range are reserved for a specific purpose. Multicast sources send traffic to a particular group address in this range which all receivers listen to. The source address in a multicast session is always a unicast IP address, and the destination address is an address from the Class D address space. Hosts interested in receiving traffic for a particular group join the group by sending a message to the directly connected router. They can also leave the group at any time.
The Layer 3 (L3) IP multicast addresses are mapped to a Layer 2 (L2) Media Access Control (MAC) address to be used for encapsulation before the actual delivery of the packets. Receivers that join a particular group listen for the corresponding MAC address to receive and process the packets. Hosts that do not join a group do not listen to the corresponding MAC address, and therefore drop those packets.
There are various protocols that work together to make IP multicasting work.
Hosts use Internet Group Management Protocol (IGMP) to inform a directly connected router about their intention to receive traffic for a particular group. IGMP messages are exchanged between the hosts and routers to determine if there are any attached receivers for a group.
IP multicast relies on a data distribution tree built by a multicast routing protocol to deliver packets from the source to the receivers when they are connected to different networks across routers. One form of the tree is called the source tree, which has its root at the source and branches forming a spanning tree through the network to the receivers. It is also referred to as a Shortest Path Tree (SPT), since traffic flows from the source to the receivers in the shortest path through the network. The other form of the tree is called the shared tree, which uses a single common root placed at some chosen point in the network. This shared root is called a Rendezvous Point (RP). Traffic flows from the source to the receivers through the RP, with a source tree from the source to the RP, and a shared tree from the RP to the receivers.
There are various multicast routing protocols that can run between the routers to form either a source or shared tree. Depending on whether multicasting is implemented within an Autonomous System (AS) or between different ASs, they are classified as intradomain or interdomain multicast routing protocols, respectively.
Some of the intradomain multicast routing protocols are Distance Vector Multicast Routing Protocol (DVMRP), Multicast Open Shortest First (MOSPF) and Protocol Independent Multicast (PIM). Cisco devices support only PIM and can interoperate with DVMRP devices. PIM can either operate in PIM Dense Mode (PIM-DM) or PIM Sparse Mode (PIM-SM). PIM-DM uses a periodic flood and prune mechanism to build source trees. PIM-SM uses an RP to build a shared tree. Once a distribution tree is built, routers use the Reverse Path Forwarding (RPF) check to forward multicast traffic from one interface to another. The RPF check verifies whether the packet arrived on the correct interface pointing towards the source to avoid loops. PIM can use the information learned though a unicast routing protocol, such as Routing Information Protocol (RIP), Enhanced Interior Gateway Routing Protocol (EIGRP), Open Shortest Path First (OSPF) and Intermediate System-to-Intermediate System (IS-IS). It can also use static routes to find the interface leading to the source.
The interdomain multicast routing protocols which can be used are Multicast Border Gateway Protocol (MBGP), Multicast Source Discovery Protocol (MSDP) and Source Specific Multicast (SSM). An intradomain multicast routing protocol such as PIM-SM is still used within the AS, but the unicast route information learned through BGP needs to be used to perform the RPF check. MBGP is an extension of ordinary Border Gateway Protocol (BGP) that can indicate whether an advertised prefix is used for unicast routing, multicast RPF checks, or both. It is useful when there are multiple paths between two different ASs, and different paths have to be used for the unicast and multicast traffic between them. MSDP is used along with PIM-SM for learning the active sources in another AS transmitting multicast traffic for a group. Since each AS will have its own RP that knows about the active sources within their AS, the RPs exchange this information using MSDP.
Another application of MSDP is Anycast RP, which can be used in a Sparse Mode network to provide RP fault tolerance and load sharing within an AS. SSM is another protocol which is an extension of the PIM protocol that allows hosts to receive traffic only from specific sources. Since receivers indicate the source from which they require traffic for a group they have joined, it eliminates the need for RP and MSDP, and builds a SPT from the source to the receiver.
Multicasting leads to unwanted packet flooding to parts of the network where there are no receivers. This occurs in LAN environments comprised of LAN switches, which is a Layer 2 (L2) device forwarding packets based on MAC addresses. There are various methods, such as IGMP snooping, the proprietary Cisco Group Management Protocol (CGMP), and Router-Port Group Management Protocol (RGMP) that avoid this unwanted flooding.
For more information on IP multicasting, refer to IP Multicast Technology Overview.
