CC*Integration-Large: Programmable Network Testbed for 400 Gbps Science DMZ

高能物理等科学驱动因素的数据存储和传输需求已经远远超过了常规的校园流量，并且还在继续增长。这种爆炸性的增长反过来又对监测和管理研究网络的任务提出了挑战。 该项目将设计，实施和部署PRONET，这是一个支持管理科学DMZ流量的前沿研究的测试平台。PRONET将实现（i）实时监控每秒太比特的流量，以执行安全和性能等任务;（ii）以网络线路速率进行复杂处理，以支持当前的科学DMZ用例（例如，检测安全攻击）和新兴攻击（例如，处理IoT流量）;以及（iii）验证可编程网络对真实世界流量的研究。创新的主要领域是：（i）设计一个系统，用于真实的实时监控科学DMZ流量，以满足操作和实验的需要，该系统可以扩展到Tbps的流量;（ii）监控科学DMZ内流量的新能力，这将使增强的安全性和新类型流量的特征化成为可能;（iii）展示可编程网络如何通过支持将功能卸载到异构和分布式可编程硬件，在科学DMZ环境中实现与安全攻击检测、基于机器学习的流量分类和物联网流量处理相关的各种新兴用例;及（iii）透过在实际网络流量上试验可编程网络原型而获得的见解，可推动该领域的研究。该小组由在网络研究及校园网络管理方面具有互补专长的专业人员组成，基础设施演进PI将与普渡的中央IT组织合作，将PRONET集成到普渡的网络基础设施中。该项目将通过展示这些环境中可编程网络的用例来影响未来的科学DMZ设计。这些经验将传播给网络研究人员以及管理科学DMZ环境的从业人员。该项目将推动可编程网络的采用，从而实现更安全、可靠和灵活的网络环境。该项目将广泛涉及研究生和本科生，PI将努力积极招募代表性不足的群体的学生。该奖项反映了NSF的法定使命，并已被认为是值得通过使用基金会的智力价值和更广泛的影响审查标准进行评估的支持。

The data storage and transfer needs of science drivers such as high energy physics far exceeds regular campus traffic already, and continues to grow. This explosive growth in turn challenges the task of monitoring and managing research networks. This project will design, implement and deploy PRONET, a testbed to support cutting-edge research for managing Science DMZ traffic. PRONET will enable (i) real-time monitoring of Terabits per second of traffic for tasks such as security and performance; (ii) enable complex processing at network line rate to support both current Science DMZ use cases (e.g., detect security attacks) and emerging ones (e.g., process IoT traffic); and (iii) validation of programmable networking research on real-world traffic.Major areas of innovation are: (i) the design of a system for monitoring Science DMZ traffic in real time for both operational and experimental needs that can scale to Tbps of traffic; (ii) new capabilities to monitor traffic within the Science DMZ which will enable both enhanced security, and characterization of new types of traffic; (iii) demonstrating how programmable networking can enable a wide range of emerging use cases in Science DMZ environments related to security attack detection, machine learning-based traffic classification, and IoT traffic processing through support for offloading functionality to heterogeneous and distributed programmable hardware; and (iii) insights obtained by experimenting with programmable networking prototypes on real-world traffic which can advance research in the area.The team comprises PIs with complementary expertise in both networking research and management of campus cyber-infrastructure. The PIs will collaborate with Purdue's central IT organization to integrate PRONET into Purdue’s network infrastructure. The project will influence future Science DMZ designs by demonstrating use cases of programmable networking in these environments. The experiences will be disseminated to both networking researchers as well as practitioners who manage Science DMZ environments. The project will drive the adoption of programmable networking, leading to more secure, reliable, and agile network environments. The project will extensively involve graduate and undergraduate students, and the PIs will work to actively recruit students from underrepresented groups.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.