Formal Definitions and Performance Comparison of Consistency Models for Parallel File Systems
Chen Wang, Kathryn Mohror, Marc Snir
TL;DR
The paper addresses the high cost of POSIX-style strong consistency in large-scale HPC I/O. It introduces a formal SCNF framework that unifies memory and storage consistency models and a layered BaseFS–CommitFS–SessionFS implementation to enable fair performance comparisons. Through a comprehensive study, it shows that weaker, properly synchronized models like session consistency can significantly boost I/O bandwidth for small, random reads (e.g., up to several-fold improvements) and improve scalability in deep-learning workloads and SCR restart. The work provides a principled path to evaluating and adopting relaxed storage models in parallel file systems, with practical implications for I/O performance and programmability.
Abstract
The semantics of HPC storage systems are defined by the consistency models to which they abide. Storage consistency models have been less studied than their counterparts in memory systems, with the exception of the POSIX standard and its strict consistency model. The use of POSIX consistency imposes a performance penalty that becomes more significant as the scale of parallel file systems increases and the access time to storage devices, such as node-local solid storage devices, decreases. While some efforts have been made to adopt relaxed storage consistency models, these models are often defined informally and ambiguously as by-products of a particular implementation. In this work, we establish a connection between memory consistency models and storage consistency models and revisit the key design choices of storage consistency models from a high-level perspective. Further, we propose a formal and unified framework for defining storage consistency models and a layered implementation that can be used to easily evaluate their relative performance for different I/O workloads. Finally, we conduct a comprehensive performance comparison of two relaxed consistency models on a range of commonly-seen parallel I/O workloads, such as checkpoint/restart of scientific applications and random reads of deep learning applications. We demonstrate that for certain I/O scenarios, a weaker consistency model can significantly improve the I/O performance. For instance, in small random reads that typically found in deep learning applications, session consistency achieved an 5x improvement in I/O bandwidth compared to commit consistency, even at small scales.