Noise in the Clouds: Influence of Network Performance Variability on Application Scalability

TL;DR

This paper investigates how network performance variability in cloud environments affects the scalability and cost-efficiency of HPC workloads, highlighting the impact of network noise on application performance.

Contribution

It provides a detailed analysis and simulation of network noise effects on HPC scalability across multiple cloud providers and on-premise systems, emphasizing the importance of network stability.

Findings

Network noise can significantly reduce HPC application performance.

Cloud network variability impacts scalability and cost.

Validation across multiple providers confirms the effects.

Abstract

Cloud computing represents an appealing opportunity for cost-effective deployment of HPC workloads on the best-fitting hardware. However, although cloud and on-premise HPC systems offer similar computational resources, their network architecture and performance may differ significantly. For example, these systems use fundamentally different network transport and routing protocols, which may introduce network noise that can eventually limit the application scaling. This work analyzes network performance, scalability, and cost of running HPC workloads on cloud systems. First, we consider latency, bandwidth, and collective communication patterns in detailed small-scale measurements, and then we simulate network performance at a larger scale. We validate our approach on four popular cloud providers and three on-premise HPC systems, showing that network (and also OS) noise can significantly impact performance and cost both at small and large scale.