Protocol design and performance evaluation for network science and engineering

在过去的半个世纪里，互联网取得了巨大的成功。网络研究界已经意识到，我们现在正处于为下半个世纪设计和评估下一代网络协议的门槛。在云计算、智能电网、网络物理系统、在线社交网络等以网络科学和工程的名义出现的推动下，已经发起了许多努力，如FIND/GENI/NetSE(美国)、FIRE(欧盟)、JGN(日本)等。除了对网络研究的性能、可靠性、安全性和可测性等传统需求外，这些新的网络范例也带来了许多新的挑战，如能效、超可伸缩性和异构性等，其中网络拓扑扮演着越来越重要的角色，并影响到所有上层协议。拟议的研究计划与加拿大的其他学术和行业研究计划一起，将使加拿大社会和经济在这场国际竞争中保持领先地位。特别是，传统的网络协议遵循存储转发的原则。随着认知无线电、压缩感知、协作通信和空白网络等物理层技术的发展，以及网内处理和覆盖网络等上层技术的发展，我们的研究计划将设计原则扩展到“存储-进位-处理-转发”，充分利用可预测的、可控制的或用户的移动性来承载信息，并处理网络内的数据流，以达到最大容量。这是研究界首次提出这一原则，尽管近几年来的文献一直在分别考察进位和过程。由于许多新兴的复杂系统，如车载自组织网络、无线传感器网络、移动社交网络和对等网络中的时间/位置变化的拓扑结构，我们认为对于我们必须依赖的关键基础设施设计和评估新的协议是非常必要的。

The Internet has achieved great success in the last half century. The networking research community has realized that we are now at the doorstep of designing and evaluating next-generation network protocols for the next half century. There are many efforts already initiated, such as FIND/GENI/NetSE (US), FIRE (EU), JGN (Japan), etc, driven by the emerging of cloud computing, smart grids, cyber-physical systems, online social networks, etc, collectively under the name of network science and engineering. Besides the traditional demand on networking research, such as performance, reliability, security and testability, these new networking paradigms have created many new challenges such as energy-efficiency, ultra-scalability and heterogeneity, and so on, among which the network topology plays an increasingly important role and affects all upper-layer protocols. The proposed research program, together with other academic and industry research programs in Canada, will keep Canadian society and economy on the leading edge of this international competition. Particularly, traditional network protocols follow the "store-and-forward" principle. With the advance of physical-layer technologies such as cognitive radio, compressive sensing, cooperative communication and white-space networking, as well as the upper-layer technologies such as in-network processing and overlay networking, the design principle is expanded to "store-carry-process-forward" in our research program, to fully utilize predictable, controlled, or user mobility to carry information and to process data flows inside the network to achieve the maximum capacity. This is the first time that such principle is proposed in the research community, although the literature has been examining carry and process separately in last few years. Due to the time/location-varying topology in many emerging, complex systems such as vehicular ad hoc networks, wireless sensor networks, mobile social networks and peer-to-peer networks, we believe such a thorough examination of store, carry, process and forward is very necessary to design and evaluate new protocols for the critical infrastructures that we have to rely on.