Research On Resource Scheduling Methods For Cloud Based Live Media Streaming System

In the next generation of Internet, along with the development of the broadbandand web2.0, IPTV, VoD, high-definition video streaming applications graduallybecome the mainstream of broadband application. Streaming media application withhigh bandwidth, high traffic and high quality of service requirements put forward ahuge challenge to the Internet which based on “do its best”.CDN as use the adoptionof static deployment of resources, it will lead to the insufficient about resourcesdeployment or waste of resources. And the using of P2P (Peer to Peer) streamingmedia distribution, due to the freedom of P2P network, it will cause the instability ofservice quality.Cloud computing as a new network computing model, it can provide users withnearly unlimited extensible Shared virtualized resources, the characteristics of theelastic computing provides the opportunity to solve the problems faced by the CDNand P2P. Streaming media distribution which building on cloud platforms, solvingthe problems existing in the traditional distribution and dynamic change of userparoxysmal requests, at the same time can significantly reduce the operation cost ofstreaming media application providers, thus receive widely attention.Constructing streaming media distribution system based on cloud platform, alarge number of highly dynamic demand facing from different geographical areas, inorder to ensure the quality of service they need to adopt to the nearest district domainmultiple cloud service provider’s data center, and they need to build a distributionnetwork for streaming media applications based on the distributed cloud data centerarchitecture (hereinafter referred to as the distributed cloud). Aiming at the typicalapplication scenario for the multi channel and multi-user area, study the establishmentof distribution network based on distributed cloud and content distribution algorithm,considering the distributed topology, the cloud data center billing methods,theposition of the user and request rate such factors, through it is modeled as a directedsteiner tree, content distribution algorithm is given to try to lower the price to meetthe requirements of the streaming media quality of service. After verification andperformance analysis under the environment of toolbox in matlab and yalmip linearprogramming, the results show that under different topologies, heuristic algorithm canget near optimal solution with low time complexity and can be easily applied in thebusiness of streaming media content distribution systems.