On Optimization Issues In Supporting High Bandwidth Multimedia Applications

尽管互联网取得了成功，其带宽过度配置，连接冗余，以及对操作的理解增加，但互联网用户仍然会遇到拥塞和服务中断。此外，网络应用越来越复杂，需要更高的可靠性和更高的带宽。特别是高带宽一对多多媒体应用的激增迫在眉睫。如果没有有效的网络设计方案，用户不可避免地会遇到偶尔的网络服务降级或故障。本研究的目的是设计，实施和评估网络架构的一对多的高带宽应用程序，通过多层次的角度实现最佳的资源利用率，性能和可靠性。 PI将开发可扩展的和有效的算法的设计和重新配置的虚拟拓扑组播，同时支持单播在光层;调查开销，可扩展性和可靠性之间的权衡在IP光网络模型;探索有效的多媒体传输方案，特别是在多供应商网络;和分析多层交互和必要的合作组播应用。 其思想价值在于，在单播应用中采用整体多层方法来支持大规模多播，从而解决实际中的优化问题。多个层的隐式或显式协作为一对多高带宽应用的网络设计提供了新的维度。优化算法的开发和创新的性能效益的量化预计将在计算机科学中具有广泛的适用性。该项目将通过接触来自代表性不足的群体的学生进行研究项目，通过将最新的研究发展融入本科生和研究生课程的教学活动，提高教育质量，并广泛传播项目成果。

Despite the success of the Internet, its bandwidth overprovisioning, redundancy in its connectivity, and increased understanding of operation, Internet users still experience congestion and service disruption. Furthermore, network applications are increasingly sophisticated requiring greater reliability and higher bandwidth. Particularly proliferation of high bandwidth one-to-many multimedia applications is imminent. Without effective network design schemes in place, users inevitably encounter occasional network service degradation or breakdowns. This research aims to design, implement, and evaluate network architecture for one-to-many high bandwidth applications through multiple layer perspectives toward achieving optimal resource usage, performance and reliability. The PI will develop scalable and efficient algorithms of design and reconfiguration of virtual topologies for multicast while supporting unicast in optical layer; investigate the trade-off among overhead, scalability, and reliability in IP over optical network models; explore efficient multimedia delivery schemes particularly over multi-provider networks; and analyze multi-layer interaction and necessary cooperation for multicast applications. The intellectual merit lies in taking a holistic multi-layer approach to large scale multicast support in the presence of unicast applications so as to address the optimization issues practically. The implicit or explicit cooperation of multiple layers provides new dimensions to network design for one-to-many high bandwidth applications. The developing of optimization algorithms and the quantifying the performance benefits of innovations are expected to have a wide applicability in computer science.The project will broadly impact the universities, industry and research community in a number of ways by reaching out to students from underrepresented groups for the research projects, enhancing the quality of education by integrating the latest research developments into instructional activities of undergraduate and graduate curriculum, and making project results available widely.