Booting Over InfiniBand for Consolidation Savings

Booting Over Infiniband for Consolidation Savings

by Frank Leers
November 2008

InfiniBand, with its high speed and low latency, is becoming the interconnect of choice for large High Performance Computing (HPC) clusters. Many clusters currently use InfiniBand for file sharing and data transfer. Configuring the cluster nodes to boot over the InfiniBand fabric is an attractive option for many environments, as Booting over InfiniBand (BoIB) can offer compelling consolidation and scaling benefits for systems with tens to thousands of compute nodes.

Contents

  • Mellanox BoIB Implementation Overview
  • Boot over InfiniBand Implementation
    • Preboot eXecution Environment (PXE)
    • DHCP
    • TFTP
  • Test Environment
    • Hardware Configuration
    • Software Components
  • Installation and Configuration
    • Flash the HCA
    • Adjust BIOS Settings
    • Patch the DHCP Server Source Code
    • Reconfigure DHCP
    • Setup Sun HPC Software, Linux Edition and oneSIS Software
    • Modify the pxelinux.cfg File
  • Future Work
  • Conclusion
  • About the Author
  • Acknowledgements
  • References
  • Ordering Sun Documents
  • Accessing Sun Documentation Online
  • Appendix: init File Output
About the Authors

Frank Leers is a Senior Staff Engineer at Sun Microsystems, currently working in the Strategic Engagements Team (SET) as a High Performance Computing Technical Specialist. In his current role, Frank is focused on deploying large HPC projects at diverse Sun customer sites, bringing together best of breed Sun hardware and cutting edge HPC software. Before Joining SET, Frank spent nearly ten years in Sun Services as an escalation resource, working as an Area Systems Support Engineer (ASSE) in the Western US. While working in the Services division, Frank partnered with Sun customers to solve complex problems across many disciplines including systems, storage, operating environments, and customer applications. Frank is an active member of several Sun communities, including the High Performance Computing ACES, Technical Systems Ambassadors, and Storage ACES.

Acknowledgements

The author would like to recognize the significant contributions made by Matthew Bohnsack during the creation of content for this Sun BluePrints article.

Rate this blueprint (Log In to vote.)
Choices Your Vote

Great

Good

Fair

Poor

Labels

new new Delete
hpc hpc Delete
x64 x64 Delete
blueprint blueprint Delete
Enter labels to add to this page:
Please wait 
Looking for a label? Just start typing.

Sign up or Log in to add a comment or watch this page.


The individuals who post here are part of the extended Sun Microsystems community and they might not be employed or in any way formally affiliated with Sun Microsystems. The opinions expressed here are their own, are not necessarily reviewed in advance by anyone but the individual authors, and neither Sun nor any other party necessarily agrees with them.

Copyright 1994-2009 Sun Microsystems, Inc.
Powered by Atlassian Confluence
Sun Guidelines on Public Discourse Privacy Policy Terms of Use Trademarks Site Map Employment Investor Relations Contact