CSR -- AES: Collaborative Research: Next Generation Internet Proxy Systems

在教育、医学研究和实践、新闻媒体以及娱乐和游戏行业等许多重要应用领域，互联网流媒体内容交付的要求越来越高。虽然代理缓存技术在基于文本的Web内容分发中已经得到了广泛而成功的应用，但由于媒体对象一般规模巨大，且对连续、及时的流要求严格，需要长时间的专用、连续带宽，因此不适合流媒体。为显著提高低成本、高质量的互联网流媒体服务质量，本项目将设计和实施下一代互联网代理系统，该系统包括高效、复杂的流媒体功能以及代理中的资源协调和供应能力。提议的项目将开发和分析一组算法，用于具有交互请求的流缓存，资源约束转码，以及不同类型缓存对象的资源配置和协调，并将通过实现核心流功能，准确预取媒体段以确保良好的流质量，以及在代理中启用格式无关的流来生成一个真正的系统。

Streaming media content delivery in the Internet has become increasingly demanding in many important application areas, such as education, medical research and practice, news media, and entertainment and gaming industries. Although the proxy caching technique has been widely and successfully used for textbased Web content delivery, it is not suitable for streaming media due to the general huge sizes of media objects and the rigorous requirements of continuous and timely streaming, which demand dedicated and continuous bandwidth for a long period of time. To significantly improve the quality of Internet streaming services with low cost and high quality, the project will design and implement Next Generation Internet Proxy Systems that include effcient and complex streaming functions and resource coordination and provisioning capability in the proxy.The proposed project will develop and analyze a set of algorithms for streaming caching with interactive requests, resource-constrained transcodings, and for resource provisioning and coordination of different types of cached objects, and will produce a real system by implementing core streaming functions, accurately prefetching media segments to ensure good streaming quality, and enabling format-independent streaming in the proxy.