A principled approach to managing routing in large isp networks

Internet Service Providers (ISPs) are the core building blocks of the Internet, and play a crucial role in keeping the Internet well-connected and stable, as well as providing services that meet the needs of other ASes (and their users). As a result, an ISP plays different roles in its operation: (1) as part of the Internet, an ISP is expected to help keep the global network stable; (2) when interacting with neighboring networks, an ISP faces diverse requirements from different neighbors about the kinds of routes they prefer; and (3) internally, an ISP needs to maintain and upgrade its own network periodically, and wants avoid disruptions during those operations as much as possible. As the Internet has become an integral part of the world’s communications infrastructure, today’s ISPs face a number of routing management challenges at these different scopes, which include: (i) maintaining the stability of the global Internet while meeting the increasingly demands for providing diverse routes from its customers, (ii) supporting more flexible routing policy configuration in bilateral contractual relationships with its neighbors, and (iii) making network maintenance and other network management operations in their own networks easier and less disruptive to routing protocols and data traffic. This dissertation takes a principled approach to addressing these challenges. We propose three abstractions that guide the design and implementation of our system solutions. First, we propose the abstraction of a “neighbor-specific route selection problem” and a corresponding “Neighbor-Specific BGP” (NS-BGP) model that capture the requirement of customized route selection for different neighbors. Since one ISP’s route selection decisions could cause the global Internet to become unstable, we prove the conditions under which the Internet is guaranteed to remain stable even if individual ISPs make the transition to this more flexible route-selection model. Second, we model policy configuration as a decision problem, which offers an abstraction that supports the reconciliation of multiple objectives. Guided by this abstraction and the Analytic Hierarchy Process, a decision-theoretic technique for balancing conflicting objectives, we designed and implemented a prototype of an extensible routing control platform (Morpheus) that enables an ISP to select routes for different neighbors individually and make flexible trade-offs among policy objectives through a simple and intuitive configuration interface. Finally, we propose the abstraction of the separation between “physical” and “logical” configurations of routers, which leads us to the design and prototype implementation of “virtual router migration” (VROOM), a new, generic technique to simplify and enable a broad range of network management tasks, from planned maintenance to reducing power consumption. Collectively, the contributions of the dissertation provide simple system solutions for an ISP to autonomously manage its routing more flexibly and effectively without affecting global routing stability.