Minimum cost broadcast routing

Is there any method where we can get a minimum cost broadcast routing scheme without using the spanning t开发者_JAVA百科ree algorithm?

Any references to guide me on this will be of great use to me.

Any algorithm for implementing a minimum cost broadcast (or multicast) routing scheme in the end boils down to constructing a least-cost spanning tree (rooted at the multicast source) of the full graph which represents the network.

There are various algorithms for computing the least-cost spanning tree.

IP multicast routing protocols such as PIM rely on least-cost spanning tree which is computed by the IGP (OSPF or ISIS) using the Dijkstra algorithm.

Older protocols, such as DVMRP, rely on a distance-vector protocol to compute the spanning tree.

One could theoretically use other algorithms to compute the least-cost spanning tree (e.g. Bellman-Ford) although I know of no implementation that does so in practice.

To add to Cayle, considering you mentioned spanning tree I assume you are not talking about layer 3 (pim, dvmrp, ospf, isis, rip, bgp, etc) and you are instead talking about layer 2. TRILL is a new standardized protocol that does nearly minimum cost broadcast routing (assuming your source is at a root of a distribution tree). IEEE 802.1aq is a protocol still be standardized that does a similar routing scheme.

Trill(Transparent Interconnection of Lots of links ) is the best option against STP.Because in stp there is a shortest dedicated path from source to destination.When Trill is used for multipathing.So using Trill we can use maximum bandwidth of the network.The research is also going on Trill.Trill usually use ISIS protocol for internal routing and Dijkstra algorithm for shortest path.

Trill basically work on L2 as well as on L3.If you are looking for implementation of trill you have to refer RFCs like 6325,6326,7177 and many more. If you are looking for switches with Trill implemented.Cisco,juniperNetwork, Huawei and many other companies implemented it partially in the switches.

Thanks