MRI: Development of Next-Generation Network Research Testbeds

该项目通过扩展软件和创新算法，建立了一组联合的本地和高度可配置的“互联网仿真器”的框架(每个异质的，在本地控制下，提供一套独特的硬件)，扩展了以前的研究，为分布式系统和网络的研究构建了一种新型的实验环境。许多机构目前使用这种时间和空间共享的设施进行研究。个人站点用户现在将拥有本地硬件的优先级，以及对本地管理、安全和资源分配策略的控制权。这种自主性有望鼓励“有机增长”，而智能分配、调度和交换系统将提供抽象。研究人员和学生可以不考虑本地的复杂性，而不是看到一大堆各种各样的硬件，包括标准PC、网络处理器、无线节点等。对于希望在“真实”互联网上运行的实验来说，仿真实验室之间的链路应该是一个有用的资源，而受控带宽链路应该屏蔽不能在互联网上运行的实验。对无线、移动节点的支持应该会为实验开辟新的途径。在完成这一任务时涉及许多挑战：创建管理扩展的软件新算法和技术的应用利用在非静态广域系统(NP-Complete Problem)上的Emulab之间的网络信息的抗故障对等共享来扩充现有数据库开发自动化系统以控制实验调度和分配，同时维护简单的用户界面扩展Emulab算法和软件以处理更广泛种类的硬件，包括无线节点开发用于在机器之间共享、保存和恢复状态的方法，同时在单个计算机上多路复用多个虚拟节点。因此，该工作将分布式系统试验台中可配置实验的现有能力概括为包括无线和有线链接，移动节点和固定节点，以及具有千兆链路的高端节点。Emulab在实验和教学中显示出巨大的个人力量；因此，将这些连接到一个大型的“研究和教育网格”中可以提供更大的能力和灵活性。获得一个极其多样化和强大的资源，小型机构可以做出贡献，因为向网格添加站点将是容易和廉价的。由于Emulab已经作为一种教育工具在几个机构中使用，其教育和研究影响仍然非常广泛。

This project, establishing the framework for a federated set of local and highly configurable "internet emulators," known as Emulabs, (each heterogeneous, under local control, offering a unique set of hardware) with expanded software and innovative algorithms, extends previous research, building a new type of experimental environment for research in distributed systems and networks. Many institutions presently use this time- and space-shared facility for research. Individual site users will now have priority for local hardware, as well as control over local administrative, security, and resource allocation policies. This autonomy is expected to encourage "organic growth," while intelligent allocation, scheduling, and swapping systems will provide abstraction. Researchers and students can be unconcerned with local complexities, instead of seeing one large collection of a wide variety of hardware, including standard PCs, network processors, wireless nodes, and more. The links between Emulabs should be a useful resource for experiments wishing to operate on the "real" Internet, while controlled-bandwidth links should shield experiments that do not. Support for wireless, mobile nodes should open new avenues for experimentation. Many challenges are involved in achieving this mission: Creating software to manage the extension Application of novel algorithms and techniques Augmenting existing database with failure resistant peer-to-peer sharing of network information between Emulabs Scheduling on a non-static, wide-area system (NP-complete problem) Developing automated systems to control experiment scheduling and allocation while maintaining simple user interfaces Extending Emulab algorithms and software to handle a wider variety of hardware, including wireless nodes Developing methods for sharing, saving and restoring state between machines Maintaining performance isolation while multiplexing multiple virtual nodes on a single computerThus, the work generalizes existing capabilities for configurable experimentation in distributed systems testbeds to include wireless as well as wired linked, mobile as well as fixed nodes, and high-end nodes with gigabit links. Emulabs exhibit great individual power in experimentation and teaching; hence, connecting these into one large "Research and Education Grid' offers more power and flexibility. Gaining access to an immensely diverse and powerful resource, small institutions can contribute, because adding sites to the Grid will be easy and cheap. Since Emulab is already in use as a tool for education in several institutions, its education and research impacts remain very broad.