CAREER: A Flexible Flow- and Error-Control Protocol-Integration Architecture for Multicast Services Over Mobile Networks

移动组播提供了一种高效而灵活的方式，将数据或信息从一个源同时传播到多个与位置无关的接收者。然而，为移动网络上的多播服务提供有效的流量和差错控制带来了在有线或单播网络中没有遇到的许多挑战。首先，组播本身在流协议和差错控制协议中都存在可伸缩性问题。其次，由于无线噪声、信道衰落和频繁切换，无线链路上的随机丢失概率比有线网络高得多。第三，无线链路引入了非拥塞损失和异质损失特性，这就需要新的流量和差错控制方案。最后，在同一网络中，单播与多播、无线有线以及实时与非实时应用的共存导致服务质量(Qos)要求的显著多样性和公平性问题。不幸的是，目前支持移动组播的新流和差错控制协议的发展没有跟上这些挑战的步伐。如果不立即投资于这一领域的研究工作，这些挑战可能会减缓我们发展下一代互联网的速度。目前还没有一种“一刀切”的协议可以通过最好地满足移动网络上多播服务施加的显著不同的流量和差错控制要求来解决所有上述挑战。为了克服这一问题，我们建议开发一种新的体系结构，该体系结构使用解耦的流和差错结构，使得不同的流和差错控制协议能够灵活地集成，以有效地满足任何给定的特定移动组播传输的服务质量要求。另一方面，提出的体系结构还允许不同的流量和差错控制协议在公共网络中共存。虽然本方案中提出的体系结构只关注应用于移动网络上的组播服务的流量和差错控制，但它是一种通用的协议开发方法，因此可以在许多其他协议开发应用中受益。该方案中需要详细探讨的具体研究课题/领域包括：开发灵活的体系结构，在其上可以有效地集成不同的流和错误协议，以满足移动网络上多播服务的不同服务质量要求。设计统一的/通用的接口技术/结构，用于灵活地集成不同类型的解耦的流和错误协议。调查多个协议的公平性和共存问题，以及它们与当前互联网中现有的TCP业务的TCP友好性。通过各种分析方法、仿真工具和协议接口技术来验证和评估所提出的体系结构、流和错误协议以及协议接口技术智能优势：提出的研究项目有望对下一代互联网中设计灵活集成移动组播和其他服务的多种流量和差错控制协议的通用体系结构的开发产生重要影响。目前正在与多家公司进行产业合作，以获得反馈，并在现实生活中测试开发结果，并使技术转移成为可能。PI与IBM T.J.Watson研究中心和TI数字视频通信部门正在进行研究合作，这将在本项目中得到利用。广泛影响：该提案的教育部分侧重于在德克萨斯农工大学建立一个强大的计算机网络本科生和研究生课程。为了实现这一目标，PI计划在未来几年内开展一系列教育项目：.开发一门新的移动无线网络研究生课程.为本科计算机网络课程设计一套实验室.整合计划中的研究原型，作为无线网络研究生课程的试验台.

Mobile multicast provides a highly efficient and flexible way of simultaneously disseminating data or information from one source to multiple location-independent receivers. However, providing the efficient flow and error-control for multicast services over the mobile networks presents many chal- lenges that are not encountered in the wired or unicast networks. First, multicast itself imposes the scalability problems in both flow and error control protocols. Second, random-loss probability over the wireless links are much higher than the wired networks due to the wireless noise, channel fad- ing, and frequent handoffs. Third, wireless link introduces non-congestion losses and heterogeneous loss properties, which requires the new flow and error-control schemes. Finally, the coexistence of unicast with multicast, wired with wireless, and real-time with non-real-time applications within a common network causes the significant diversity of Quality of Services (QoS) requirements and fair- ness issues. Unfortunately, the current advances in the development of new flow and error-control protocols supporting mobile multicast are not keeping pace with these challenges. Without investing immediate research effort in this area, these challenges will likely slow the rate at which we develop the next-generation Internet.There is no "one-size fits all" protocol that can solve all the above challenges by best meeting the significantly diverse flow and error-control requirements imposed by multicast services over mobile networks. To overcome this problem, we propose to develop a novel architecture which, using the decoupled flow and error structure, enables flexible integrations of different flow and error control protocols to efficiently fit the QoS requirements for any given specific mobile multicast transport. On the other hand, the proposed architecture also allows different flow and error-control protocols to coexist in a common network. While the proposed architecture focuses only on the flow and error control applied to multicast services over mobile networks in this proposal, it is a generic protocol developing methodology and thus can be beneficial in many other applications of protocol development. The specific research topics/areaB to be explored in detail in this proposal include:.Develop the flexible architecture on which different flow and error protocols can be integrated efficiently to meet the different QoS requirements for multicast services over mobile networks..Design the uniform/general interfacing techniques/structures for flexible integrations of different types of decoupled flow and error protocols..Investigate the fairness and coexistence issues of multiple protocols and their TCP-friendliness with currently existing TCP-traffics in the Internet..Validate and evaluate the proposed architecture, flow and error protocols, and protocol interfacing techniques through various analytical approaches, simulation tools, and testbed systems.Intellectual merit: The proposed research project is expected to have an important impact on the development of the general-purpose architecture for designing flexible integrations of multiple flow and error-control protocols for mobile multicast and other services in the next-generation Inter- net. The industrial collaboration is being pursued with a number of companies to obtain feedback and test the developed results in a real-life setting, and enable the technology transfer. The PI has ongoing research collaboration with IBM T. J. Watson Research Center and TI Digital Video Communication Department that will be leveraged upon in this project.Broader impact: The education component of the proposal focuses on building a strong un- dergraduate and graduate program in Computer Networking at Texas A&M University. To achieve this goal, the PI plans to undertake a number of educational projects during the next few years:.Develop a new graduate level course on mobile wireless networks..Design a lab set for the undergraduate course on Computer Networking..Integrate the planned research prototype as a test bed for graduate course on wireless networks.