Combined Path and Server Selection in Dynamic Multimedia Environments

This paper deals with a framework in which scheduling decisions are based on path as well as server qualities. Specifically, we address two problems in this paper. 1. Scheduling a request from a client with QoS constraints on (a) the path quality and the (b) server quality. This is framed as an optimization problem on the overall network and server loads while scheduling the request. 2. Information Collection to capture the current system state. We develop a model and algorithm for the parameter collection process that maximizes accuracy and minimizes traffic overhead. We present an algorithm to combine the two problems into one optimization problem within a unified framework. The framework has two components. The first component implements the optimized scheduling algorithm, while the second component collects the network and server parameters used in the algorithm. Our design objectives are: 1. Maximize the client request success probability. 2. Maximize the network and server utilization using load balancing techniques. 3. Tolerate the imprecision in network and server parameters 4. Minimize the overhead cost. 2. THE MODE