CAREER: Othello Hashing and Its Applications to Scalable and Dynamic Network Forwarding and Functions

在限制基础设施成本的同时为社区和企业提供高速互联网服务一直是网络协议设计的长期目标。在保持低网络成本的同时实现快速、可扩展和细粒度的网络处理方面存在挑战，这主要是由于网络设备（如快速存储器）上的资源有限。 利用可编程网络的有前途的方法，如软件定义的网络，被认为是难以在传统的网络设备上执行。该项目旨在使用一种创新的散列方案，奥塞洛，通过探索快速，内存效率高，可移植的原语网络处理的基础上创新的数据结构和算法，以提高网络性能，而无需特殊和昂贵的硬件。本计画提出奥塞洛杂凑，一种在最小完美杂凑理论基础上发展出来的键-值查询演算法。与现有的网络查找方法相比，奥塞洛哈希法实现了更快的查找速度和更小的内存成本。它利用网络可编程性来支持其查找结构的动态更新。这个项目计划使用奥塞洛开发一些重要的网络原语，包括转发信息库、软件负载均衡器、分布式数据放置和私有数据访问。这个项目的成功将证明奥塞洛哈希作为一个基本的工具，在设计新的网络算法，协议和系统，现有的工具可能不适合。奥塞洛散列的影响可能超出网络研究：PI与基因组生物学研究人员合作，将奥塞洛应用于宏基因组序列分类，取得了可喜的成果。该项目还为学生培训、本科生研究参与、多样性促进和外展活动提供了一些教育活动。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Providing high-speed Internet services to communities and businesses while limiting the infrastructure cost has been a long-term goal of network protocol designs. There have been challenges in achieving fast, scalable, and fine-grained network processing while maintaining low network cost, mainly due to the limited resources on network devices such as fast memory. The promising approach of exploiting programmable networks, such as software defined networking, were considered difficult to execute on conventional network devices. This project aims to use an innovative hashing scheme, Othello, to improve network performance without special and expensive hardware by exploring fast, memory-efficient, and portable primitives for network processing based on innovative data structures and algorithms. This project proposes Othello Hashing, a key-value lookup algorithm developed on the theoretical foundation of Minimal Perfect Hashing. Othello Hashing achieves faster lookup speed and much smaller memory cost compared to existing network lookup methods. It utilizes network programmability to support dynamic updates on its lookup structures. This project plans to develop a number of important network primitives using Othello, including forwarding information bases, software load balancers, distributed data placement, and private data access. The success of this project will demonstrate Othello Hashing as a fundamental tool in designing novel network algorithms, protocols, and systems, for which existing tools may not be suitable. The impact of Othello Hashing may go beyond the networking research: the collaboration of the PI with genome biology researchers on applying Othello to metagenomic sequence classification has led to promising results. The project also offers a number of education activities for student training, undergraduate research participation, diversity promotion, and outreach activities.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.