SafeMPI - Extending MPI for Byzantine Error Detection on Parallel Clusters

TL;DR

SafeMPI extends the MPI standard to detect Byzantine errors in cluster computing environments, enabling more robust failure recovery beyond traditional fail-stop error models.

Contribution

It introduces a novel extension to MPI that can detect Byzantine errors, addressing a gap in existing failure detection methods for high-performance clusters.

Findings

Enables detection of Byzantine errors in MPI-based clusters

Improves failure recovery capabilities for general error models

Enhances robustness of high-performance computing systems

Abstract

Modern high-performance computing relies heavily on the use of commodity processors arranged together in clusters. These clusters consist of individual nodes (typically off-the-shelf single or dual processor machines) connected together with a high speed interconnect. Using cluster computation has many benefits, but also carries the liability of being failure prone due to the sheer number of components involved. Many effective solutions have been proposed to aid failure recovery in clusters, their one significant downside being the failure models they support. Most of the work in the area has focused on detecting and correcting fail-stop errors. We propose a system that will also detect more general error models, such as Byzantine errors, thus allowing existing failure recovery methods to handle them correctly.