Designing Provably Performant Networked Systems

As networked systems become critical infrastructure, theirdesign must reflect their new societal role. Today, we build systemswith hundreds of heuristics but often do not understand their inherent and emergent behaviors. I will present a set of tools and techniques to prove performance properties of heuristics running in real-world conditions. Rigorous proofs 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. I 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. I will also show that current end-to-end congestion control algorithms that bound delay cannot avoid starvation and present a method to beamform wireless signals using thousands of antennas.