CRI: IAD: Programmable Network Infrastructure with Emerging Technologies

拟议的项目旨在建立一个可编程的网络基础设施，以促进马萨诸塞大学洛厄尔分校计算机网络和无线通信技术领域的研究和教育。这种基础设施提供了一个可扩展的网络测试平台，这在网络研究和教育中是非常理想的，但还不能从传统的现成网络设备中获得。该项目的目标是为合作研究和教育建立一个可编程的网络平台。特别是，本研究将(1)构建一个具有可编程组件的网络基础设施，用于实验新的网络设计；(2)研究基础架构的可编程性，评估其在网络处理中的性能；(3)将基础设施应用于广泛的研究课题，包括网络安全、网络处理器设计和动态频谱共享；(4)将网络与通信的实验、理论和计算研究整合到教学模块中。提议的可编程网络基础设施是一种定制的网络基础设施，配备高速有线交换机、具有可编程无线电接口的无线路由器和用于物理层测量的频谱分析仪。网络处理器、fpga和软件无线电的结合在网络的每一层和每个设备内提供了可编程性，使得在分组处理工作负载表征、安全协议评估、动态频谱访问和切片网络服务方面进行实验成为可能。这样的实验研究为理论结果提供了必要的验证，并引导了新的研究方向。最重要的是，拟议的基础设施将促进共同项目方之间的合作，以研究对GENI架构的主要研究挑战。

The proposed project aims to build a programmable network infrastructure for advancing research and education in areas of computer network and wireless communication technologies at the University of Massachusetts Lowell. This infrastructure provides an extensible network test bed that is highly desirable in network research and education, but not yet available from conventional off-the-shelf network devices.The objective of the project is to build a programmable network platform for collaborative research and education. In particular, this research will (1) construct a network infrastructure with programmable components for experimenting new network designs; (2) study the programmability of the infrastructure and evaluate its performance on network processing; (3) apply the infrastructure in a broad range of research topics, including network security, network processor design, and dynamic spectrum sharing; (4) integrate experiment, theory and computational research in networks and communications into educational modules.The proposed programmable network infrastructure is a customized network infrastructure equipped with high-speed wired switches, wireless routers with programmable radio interfaces, and spectrum analyzers for physical layer measurements. The incorporation of network processors, FPGAs, and software radios provides programmability at each layer of the network and within each device, making it possible to conduct experiments on packet-processing workload characterization, security protocol evaluation, dynamic spectrum access, and sliced network services. Such experimental research provides necessary validation to theoretical results and leads to new research directions. Most importantly, the proposed infrastructure will foster collaborations among co-PIs to study principal research challenges towards the GENI architecture.