Streaming Media Server Capacity Benchmarks and Performance Model

Service provider which provides streaming media needs to know how to measure the capacity of a streaming media server and how to evaluate real time system load. This paper proposes a set of benchmarks for measuring the capacities of streaming media servers when the media contents are encoded at variable bit rates. The result of the benchmarks shows the capacities of the servers are quite different while servers provide different quality and types of service. A workload-aware performance model of streaming media server is developed. This paper validates this model by comparing the predicted and measured media server capacities for a set of workloads.