This chapter presents a method for performance analysis and configuration of audio/video-on-demand services. These kind of services makes use of an important number of resources, requires a constant quality of service and contents, and usually have important production costs. To maintain a good quality of service and at the same time, to make a profit for the content provider, services must have the optimum configuration. With this aim, the configuration process must be based on an accurate service behavioural analysis which evaluates the quality and the quantity of resources, contents and subscribers. This analysis can be performed using monitored information extracted from servers, proxies and network monitors, and predictions of a near future behaviour using laboratory experiments. To systematize both analysis and configuration, a method must be developed in order to help service managers attain a good performance and at the same time, make a profit for their companies. All this systematic method goals are based on the principle that “a satisfied client provides more profit”. Other elements such as data sources, input data, and initial configuration should also be included. Moreover, the methodology is prepared to be extensible, and adaptable to new configuration possibilities, data, analyses, or goals.
In spite of the complexity and relative immaturity of streaming technology, a wide range of deployments of audio and video services has been developed in recent years.
A lot of organizations, from digital newspapers to public companies, are interested in these types of services in order to obtain a new way of attracting the attention of a wider audience. Nevertheless, the special characteristics of streaming services, such as the delivery of continuous information, the high consumption of resources and the need for stable conditions during the delivery of contents make them very sensitive to channel errors and delays. For these reasons several works have addressed the detailed analysis of streaming traffic over different technologies (Louginov & Radha, 2002), (Cranley & Davis, 2005), (Guo, Tan, Chen, Xia, Spatscheck, Zhang, 2006) and (Chung & Claypool, 2006), presenting interesting results from the point of view of network operators and the management of streaming flows.