| Περίληψη: | The rapid development of communications during past decades has made Internet connection available to a vast amount of users which use it for a wide variety of services. Video is one of the most popular services, currently generating the majority of the global Internet traffic and shows growing tendencies for the foreseeable future. As a result, major video-on-demand service providers (i.e. YouTube) have seen their bandwidth costs increase and will inevitably continue to do so. This demand for excessive bandwidth is caused by the simultaneous video watching of large numbers of users.
Adopting some peer-to-peer architecture can be a solution to this problem, as users can contribute their resources towards offloading the providers. Indeed, a number of research efforts the past decade have focused on providing live streaming and video-on-demand (VoD) services using peer-to-peer (P2P) architectures. However in these proposed architectures, the dynamic behavior of users has not been sufficiently studied. In a real life scenario, where users can arbitrarily decide to perform a VCR function (stop, fast forward, seeking), ignoring this behavior can significantly deteriorate the system’s efficiency and the perceived QualityofExperience (QoE). Additionally the proposed architectures are designed without taking into account the users’ heterogeneous resources (bandwidth and battery), thus leading to heavily unbalanced systems that fail to utilize these resources. This can severely harm system’s efficiency especially in cases where a significant number of users are mobile nodes with scarce battery resources.
In this dissertation, the research area of P2P VoD systems was studied and a distributed system that handles peers’ dynamic behavior and utilizes their resources was developed. Specifically our system is a set of algorithms that optimize the P2P overlay network dynamically and in a distributed fashion, making it adaptive to users’ dynamic behavior and resource changes. The available resources of the participating users are optimally exploited, keeping both the contribution from the media server(s) and the energy consumption to minimal levels. To illustrate the performance of the proposed algorithms, we have developed an extensive P2P VoD simulator that shows the efficiency, scalability and stability of our system under variant and dynamic conditions.
|
|---|
