Understanding the impact of diverse streaming workloads on end-user quality of service

Abstract

Streaming media has experienced explosive growth over the last few years and will continue to increase in popularity as individual users can easily produce digital images and video. Large scale streaming will require network operators to deploy and manage a larger number of streaming servers within their networks. In order to determine the number of servers needed, and to design the appropriate management policies for a particular workload, administrators will need to understand the performance characteristics of those servers, how to measure performance, and how different workloads can affect performance. In this paper we make the case for a more complete measurement methodology by showing that streaming server performance can vary substantially based on the type of streaming workload, and, as a result, affect the end-user quality of service.