Hi Akber
Not entirely sure what you are asking. If you are asking about why you would want to use multicast within your network it is primarily to do with effecient use of your bandwidth eg.
One server needs to send a video stream to multiple clients. You can do one of 3 things
1) Unicast the packet to each individual client. This means if there are 100 clients a 100 copies of each packet would need to be transmitted from the server. Not a very efficient use of bandwidth.
2) Broadcast packet to all clients. This gets you round the problem of having to transmit 100 copies of the same packet but it means that every single machine has to process the packet whether it wants to receive the packet or not. Again not very efficient.
3) Multicast. A single packet is sent and then only sent to clients that have registered an interest in receiving that packet. This means that whole portions of your network would remain untouched (unlike broadcasting) if there were no clients in that part of the network that wanted to receive the packets.
On Cisco routers you can use PIM dense or PIM sparse mode which propogate multicast routes across your network in much the same way as your IGP eg. OSPF/EIGRP propogates routes. The are differences between how routing loops are prevented etc. but at a high level this is a good enough overview.
On Cisco switches the 2 most common methods are IGMP snooping and CGMP. IGMP is the protocol used for clients to register their interest in receiving a particular stream of data. Switches can listen in on these requests so that they know which ports are associated with which multicast address(es) thus avoiding having to flood the traffic out of all ports.
CGMP is software based and works between the switch and the router to achieve pretty much the same thing ie. mapping multicast groups to ports that want to receive the traffic.
That is a very high level overview of multicast. If this hasn't answered your question or there is something in particular you were interested please come back.
Jon
