CAREER: Resource Management in Distributed Multimedia Systems

这个建议的目的是推进国家的最先进的分布式多媒体系统在研究和教育。提出了一种新的管理共享分布式多媒体网络中异构资源的方法，称为多级相对服务质量。 研究议程集中在这种方法的探索，交通控制算法的设计和评估，以及它们的实施和实证评估在一个试验台网络。 教育议程的重点是利用多媒体技术革新教育。 研究和教育议程将： 1)开发新的方法，用于在可以在共享网络环境中操作的分布式多媒体系统中的资源管理。提出了一种新的服务模型--多级相对服务质量模型。与现有的多媒体网络服务模型不同，该服务模型能够满足大多数对延迟敏感的多媒体应用的服务需求，并且能够在现有的共享网络资源基础设施中实现。 2)设计和分析了多媒体网络中具有多类相对服务质量的资源管理的新型流量控制算法。建议的研究包括开发和评估的流量控制方法相对的服务质量，方法来科普动态变化的资源管理策略，和新的路由算法。 3)设计，实现和实验评估分布式多媒体系统中资源管理的新协议机制。通过将这些机制实现为对当前使用的网络协议的扩展，现有的网络可以逐渐演进到具有服务质量保证的环境。 4)深入了解并评估交互式多媒体在教育中的使用。 一个交互式远程辅导工具，已经部分实现了由PI，使教师和学生能够通过校园网络进行互动，将提供一个实验平台，研究使用交互式多媒体工具进行教学。 5)促进将“多媒体”确立为计算机科学教育的一个主要领域。这包括开发一个面向项目的多媒体系统高级课程，使用该机构现有的多媒体教室。

The objective of this proposal is to advance the state-of-the-art of distributed multimedia systems in both research and education. A novel approach for managing heterogeneous resources in a shared distributed multimedia network is presented, called multiclass relative Quality-of Service. The research agenda concentrates on the exploration od this approach, the design and evaluation of traffic control algorithms, and their implementation and empirical evaluation in a testbed network. The education agenda is focused on the use of multimedia technology to innovate education. The research and education agenda will: 1) develop new approaches for resource management in distributed multimedia systems that can operate in a shared network environment. A new service model, called multi-class relative Quality-of-Service, is proposed. Different from existing service models for multimedia networks, the proposed service model can satisfy service requirements of most delay-sensitive multimedia applications, yet can be realized in the current infrastructure of shared network resources. 2) design and analyze novel traffic control algorithms for resource management in multimedia networks with multi-class relative Quality of-Service. The proposed research includes development and evaluation of flow control methods for relative Quality-of-Service, approaches to cope with dynamically changing resource management policies, and new routing algorithms. 3) design, implement, and experimentally evaluate new protocol mechanisms for resource management in distributed multimedia systems. By implementing the mechanisms as extensions to currently used network protocols, existing networks can gradually evolve to an environment with Quality-of-Service guarantees. 4) gain insight in and evaluate the use of interactive multimedia for education. An interactive tele-tutoring tool, already partially realized by the PI, that enables teachers and students to interact over a campus network, will provide an experimental platform to study the use of interactive multimedia tools for teaching. 5) contribute to the establishment of `multimedia' as a main area of an education in computer science. This includes the development of a project-oriented senior level course in multimedia systems that uses multimedia classrooms available at the institution.