Fair Scheduling in High Performance Computing Environments, 1st ed. 2019

This book introduces a new scheduler to fairly and efficiently distribute system resources to many users of varying usage patterns compete for them in large shared computing environments.  The Rawlsian Fair scheduler developed for this effort is shown to boost performance while reducing delay in high performance computing workloads of certain types including the following four types examined in this book:

i.        Class A ? similar but complementary workloads

ii.      Class B ? similar but steady vs intermittent workloads

iii.    Class C ? Large vs small workloads

iv.    Class D ? Large vs noise-like workloads

This new scheduler achieves short-term fairness for small timescale demanding rapid response to varying workloads and usage profiles.  Rawlsian Fair scheduler is shown to consistently benefit workload Classes C and D while it only benefits Classes A and B workloads where they become disproportionate as the number of users increases.

A simulation framework, dSim, simulates the new Rawlsian Fair scheduling mechanism. The dSim helps achieve instantaneous fairness in High Performance Computing environments, effective utilization of computing resources, and user satisfaction through the Rawlsian Fair scheduler.

With the prevalence of shared computing environments (cloud for example), how a shared resource (such as a computer/CPU/Core) is distributed and time-shared across many users will be critically important.

Current methods and algorithms used to divide resources will simply not work in large shared computing environments. For a large organization such as a bank, or pharma, the method introduced in this book will allow them to create internal shared services hat will scale to the enterprise.

This book applies behavioral economics to computing – which opens the door for many applications across the technology sector to do the same.