Maximizing the Availability and Reliability of Videos in VoD System Using Markov Chain

In a Video on Demand system, by forwarding the server video stream client by client, a chain based scheme is formed, which reduces the server load. The existing chaining schemes are not reliable because there are frequent breaks in the chain among the clients due to stringent bandwidth and buffer, stop fail, and hardware and software failures. This leads to stream failures within the client, in turn affects other clients in the cluster. Then, the performance of the overall system degrades. To achieve system reliability we formulated a continuous time Markov chain reliability model and analyzed the availability and reliability of the streams during the intricate circumstances. Numerical results show that for 5000 streams, we obtain the reliability between 60% and 80%, even though the individual clients are highly unreliable.