Networking Design for HPC and AI on IBM Power Systems

An IBM Redpaper publication

Published 26 April 2018

cover image

ISBN-10: 0738456837
ISBN-13: 9780738456836
IBM Form #: REDP-5478-00
(62 pages)

Authors: Scott Vetter, Tobias Elpelt, Rico Franke, Yanil Zeledón Miranda

Abstract

This publication provides information about networking design for IBM® High Performance Computing (HPC) and AI for Power Systems™.

This paper will help you understand the basic requirements when designing a solution, the components in an infrastructure for HPC and AI Systems, the designing of interconnect and data networks with use cases based in real life scenarios, the administration and the Out-Of-Band management networks.

We cover all the necessary requirements, provide a good understanding of the technology and include examples for small, medium and large cluster environments. This paper is intended for IT architects, system designers, data center planners, and system administrators who must design or provide a solution for the infrastructure of a HPC cluster.

Table of contents

Chapter 1. Understanding the requirements
Chapter 2. Infrastructure for high-performance computing and artificial intelligence systems
Chapter 3. Designing a solution

Follow IBM Redbooks

Follow IBM Redbooks