CARAT: Client-Side Adaptive RPC and Cache Co-Tuning for Parallel File Systems

TL;DR

CARAT is an ML-guided framework that enables scalable, online, client-side tuning of RPC and cache parameters in parallel file systems, significantly improving performance in dynamic HPC environments.

Contribution

It introduces a novel, scalable online tuning approach for PFS that operates independently on each client using only local metrics, unlike prior global or pattern-dependent methods.

Findings

Achieves up to 3x performance improvement over static configurations.

Demonstrates effectiveness across diverse I/O patterns and workloads.

Proves scalability and lightweight nature suitable for deployment.

Abstract

Tuning parallel file system in High-Performance Computing (HPC) systems remains challenging due to the complex I/O paths, diverse I/O patterns, and dynamic system conditions. While existing autotuning frameworks have shown promising results in tuning PFS parameters based on applications' I/O patterns, they lack scalability, adaptivity, and the ability to operate online. In this work, focusing on scalable online tuning, we present CARAT, an ML-guided framework to co-tune client-side RPC and caching parameters of PFS, leveraging only locally observable metrics. Unlike global or pattern-dependent approaches, CARAT enables each client to make independent and intelligent tuning decisions online, responding to real-time changes in both application I/O behaviors and system states. We then prototyped CARAT using Lustre and evaluated it extensively across dynamic I/O patterns, real-world HPC workloads, and multi-client deployments. The results demonstrated that CARAT can achieve up to 3x performance improvement over the default or static configurations, validating the effectiveness and generality of our approach. Due to its scalability and lightweight, we believe CARAT has the potential to be widely deployed into existing PFS and benefit various data-intensive applications.