Research On Scalable Application Level Multicast

Multicast enables efficient group communication by allowing the sender to transmit a single copy of data to a group of receivers making copies where necessary.In this thesis, an Application Level Multicast (ALM) solution is presented. With this approach, an overlay network is built on top of available network services and packets are multicasted at the application layer. The overlay is organized such that each end host participating in a multicast communication re-sends multicasted messages to some of its peers, but not all of them. Thus multicasting is achieved without being dependent on specific network conditions or specific network equipment maintaining multicast state information. Scalability is identified as one of the important problems hindering progress in Application Level Multicast (ALM) research and therefore the protocol proposed is capable of scaling to large group sizes. In our protocol, all nodes interested in participating in the multicast, form a group from which a mesh is created thereafter an algorithm is run on top of the mesh to build a tree which plays a major role in delivery of data. All group members keep a list of all other members in the group and from time to time exchange group membership information with the neighbors. When the number of nodes increases to thousands, the update messages cause a big overhead thereby hindering scalability. At this point hierarchies are formed. This also involves splitting of the big groups into smaller, manageable ones and thus scalability is attained.