Balanced Fair Resource Sharing in Computer Clusters

TL;DR

This paper models computer clusters as bipartite queueing systems, demonstrating quasi-reversibility, and introduces a scheduling algorithm that ensures balanced fair sharing of resources among jobs.

Contribution

It presents a novel queueing model for clusters with arbitrary compatibilities and designs a scheduling algorithm based on quasi-reversibility for fair resource sharing.

Findings

Queueing model is quasi-reversible

Scheduling algorithm achieves balanced fairness

Supports arbitrary bipartite compatibility graphs

Abstract

We represent a computer cluster as a multi-server queue with some arbitrary bipartite graph of compatibilities between jobs and servers. Each server processes its jobs sequentially in FCFS order. The service rate of a job at any given time is the sum of the service rates of all servers processing this job. We show that the corresponding queue is quasi-reversible and use this property to design a scheduling algorithm achieving balanced fair sharing of the service capacity.