Verifying the Performance of Network Control Algorithms

Abstract

As networked systems become critical infrastructure, their design must reflect their new societal role. Today, we build systems with hundreds of heuristics but often do not understand their inherent and emergent behaviors. This dissertation presents, performance verification, a set of tools and techniques to prove performance properties of heuristics running in real-world conditions. It provides an alternative to queuing and control theory, which are typically too optimistic about performance because of their limited capacity to accurately model real-world phenomena. Overly optimistic analysis can lead to heuristic designs that fail in unexpected ways upon deployment. Rigorous proofs on the other hand, can not only inspire confidence in our designs, but also give counter-intuitive insights about their performance.

A key theme in our approach is to model uncertainty in systems using non-random, non-deterministic objects that cover a wide range of possible behaviors under a single abstraction. Such models allow us to analyze complex system behaviors using automated reasoning techniques. We will present automated tools to analyze congestion control and process scheduling algorithms. These tools prove performance properties and find counter-examples where widely deployed heuristics fail. We will also prove that current end-to-end congestion control algorithms that bound delay cannot avoid starvation.