KML: Using Machine Learning to Improve Storage Systems

TL;DR

KML introduces a machine learning-based architecture integrated into operating systems to dynamically optimize storage parameters, significantly improving I/O throughput with minimal overhead across diverse workloads.

Contribution

This paper presents the novel KML architecture that replaces manual heuristics with machine learning for dynamic storage system optimization in OSs.

Findings

KML improves I/O throughput by up to 2.3x and 15x in case studies.

KML consumes less than 4KB kernel memory and has CPU overhead below 0.2%.

KML effectively adapts to complex, unseen workloads across different storage devices.

Abstract

Operating systems include many heuristic algorithms designed to improve overall storage performance and throughput. Because such heuristics cannot work well for all conditions and workloads, system designers resorted to exposing numerous tunable parameters to users -- thus burdening users with continually optimizing their own storage systems and applications. Storage systems are usually responsible for most latency in I/O-heavy applications, so even a small latency improvement can be significant. Machine learning (ML) techniques promise to learn patterns, generalize from them, and enable optimal solutions that adapt to changing workloads. We propose that ML solutions become a first-class component in OSs and replace manual heuristics to optimize storage systems dynamically. In this paper, we describe our proposed ML architecture, called KML. We developed a prototype KML architecture and applied it to two case studies: optimizing readahead and NFS read-size values. Our experiments show that KML consumes less than 4KB of dynamic kernel memory, has a CPU overhead smaller than 0.2%, and yet can learn patterns and improve I/O throughput by as much as 2.3x and 15x for two case studies -- even for complex, never-seen-before, concurrently running mixed workloads on different storage devices.