Prediction-Guided Control in Data Center Networks

TL;DR

Polyphony is a system that enables data center network operators to predict and control network quality of service, reducing tail latency events within minutes by dynamically adjusting configurations based on workload predictions.

Contribution

This paper introduces Polyphony, a novel system that combines workload monitoring, counterfactual prediction, and closed-loop control to adapt network configurations in real-time, outperforming prior static and model-free methods.

Findings

Polyphony converges to network SLOs within ten minutes.

It re-stabilizes after workload shifts within fifteen minutes.

Polyphony outperforms prior state-of-the-art methods in dynamic adaptation.

Abstract

In this paper, we design, implement, and evaluate Polyphony, a system to give network operators a new way to control and reduce the frequency of poor tail latency events in multi-class data center networks, on the time scale of minutes. Polyphony is designed to be complementary to other adaptive mechanisms like congestion control and traffic engineering, but targets different aspects of network operation that have previously been considered static. By contrast to Polyphony, prior model-free optimization methods work best when there are only a few relevant degrees of freedom and where workloads and measurements are stable, assumptions not present in modern data center networks. Polyphony develops novel methods for measuring, predicting, and controlling network quality of service metrics for a dynamically changing workload. First, we monitor and aggregate workloads on a network-wide basis; we use the result as input to an approximate counterfactual prediction engine that estimates the effect of potential network configuration changes on network quality of service; we apply the best candidate and repeat in a closed-loop manner aimed at rapidly and stably converging to a configuration that meets operator goals. Using CloudLab on a simple topology, we observe that Polyphony converges to tight SLOs within ten minutes, and re-stabilizes after large workload shifts within fifteen minutes, while the prior state of the art fails to adapt.