Modeling QoS aware VoD systems with semiotic agents

Video-on-demand systems (VoD) have been built with proprietary technologies with demanding hardware and out of reach for normal users. With the convergence of Web and multimedia technologies, it is no longer uncommon to blend streaming video and VoD into various types of applications. Building VoD systems poses challenges in quality-of-service (QoS) control for both the network and servers. VoD systems are therefore excellent case studies for context aware multimedia applications. This paper proposes a modeling framework based on semiotic agents as a methodology for design and implementation of VoD systems. Flexibility of the framework enables designers to specify rules to make VoD systems intelligent enough to handle adaptively application, work load and QoS requirements. As a proof of concept, a prototype VoD system built with distributed multiple servers had been modeled and implemented in a test bed using the semiotic agent framework. Experiments were conducted to validate the framework.