CNS Core: Medium: Towards Logically Centralized Interdomain Routing Control

点击翻译按钮获取中文摘要

Interdomain routing is the top-level protocol that holds the entire Internet together. Fundamental improvements to its robustness, security, and manageability will benefit anyone who relies on the Internet today. The Border Gateway Protocol (BGP), the Internet's standard inter-provider routing protocol, has been widely recognized as suffering from a number of fundamental problems including intermittent routing unavailability; security problems; management headaches; and most importantly difficulty evolving to meet these challenges. This project seeks to address such long-standing problems by offloading BGP's work to the cloud by enabling 'interdomain routing as a service', a radically different approach compared to today's practice. The goal of this project is to develop an interdomain routing infrastructure that is capable of incorporating de facto interdomain routing capabilities as well as enhancements proposed by researchers or practitioners to address BGP's long-standing problems. The proposed approach is a logically centralized interdomain routing architecture that enables operators to offload BGP-style route computation to a cloud service. Logical centralization of control and management has been enormously successful for enterprise networks as evidenced by the success of software-defined networking (SDN) approaches; however, interdomain routing today is largely untouched by the transformative SDN trend. The project will conduct research on the design, analysis, and implementation and evaluation of CIRCA -- a logically centralized interdomain routing architecture -- that represents a fundamental paradigm shift in the way BGP is implemented and managed today by offloading the work of "thin client" routers to the cloud. The technical research thrusts in the project include demonstrating the feasibility of the CIRCA approach at Internet scales and quantifying benefits such as significant reduction in convergence delays, improved end-to-end reachability, more expressive routing policies, what-if analysis capabilities, and improved security against control plane attacks. The project will develop a systematic open-source prototype and engage the research community and network operators through workshops and forums.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.