What Does Protocol-Independent Multicast (PIM) Mean?
Protocol-independent multicast is a family of protocols looking after the different modes of internet multicasting for successful transmission of information in one to many, and many to many modes. All protocol-independent multicast protocols have a similar format for control message. Using the routing information available with the help of the different communication protocols, protocol-independent multicast can function without being dependent on any specific routing protocol. In other words, protocol-independent multicast does not use its own mechanism of topology discovery.
Techopedia Explains Protocol-Independent Multicast (PIM)
There are four modes in protocol-independent multicast, namely:
Sparse mode: This protocol makes use of the assumption that in a multicast group, all the receivers will be sparsely distributed in the environment. It is largely for wide area usage. The protocol supports the usage of shared trees, which are nothing but multicast distribution trees rooted at a specific node. It also supports usage of source based trees, which has a separate multicast distribution tree for every source transmitting data to a multicast group. In sparse mode, its important to have a mechanism to discover the root node or rendezvous point.
Dense mode: This protocol makes the opposite assumption of the sparse mode. It assumes that in a multicast group, all receivers are densely distributed in the environment. By flooding the multicast traffic, it builds the shortest path trees and also prunes back on the tree branches when there are no presence of receivers. The protocol is based on only source based trees and as a result does not depend on rendezvous points, unlike sparse mode. This makes the dense mode easier to implement and deploy. However, the scaling property of the dense mode is poor.
Source-specific multicast: This protocol focuses on just one node which acts as a root and the trees are built based on the same. It offers a reliable, scalable and secure model for broadcasting information.
Bidirectional protocol independent multicast: It is similar to sparse mode, with the difference being in the method of data transmission. In bidirectional, the data flow is bidirectional, i.e data flows through both directions in a branch of a tree. The data is not encapsulated. Again, bidirectional does not use source based trees at all and also there is no designated router in the case of bidirectional protocol. The protocol has great scalable properties especially when there are a large set of sources for each group.