Scientific Working Group

Purpose

The Scientific Working Group is dedicated to representing and advancing the use-cases and needs of research and high-performance computing atop OpenStack. It’s also a great forum for cross-institutional collaboration. If you are (or would like to) run OpenStack to support researchers/scientists/academics and/or HPC/HTC, then please join!

Objectives

• HPC/HTC Infrastructure
• Research Data Infrastructure
• Application Infrastructure
• Social Infrastructure
• Create opportunities for the scientific community to engage the wider OpenStack community, i.e. industry

Communication

The default communication of our members is via the following open community mailing lists:

• user-committee@lists.openstack.org for coordination of working group activities. Please prefix email subject lines with the tag “[scientific-wg]”.
• openstack-operators@lists.openstack.org for operational discussion of scientific OpenStack deployment issues. Please prefix email subject lines with the tag “[scientific]”.

Please use the hashtag “scientific-wg” for tagging any etherpad URLs, code, blogs posts, scientific research and/or other social publishing platforms.

The Working Group also maintains a wiki page.

The Working Group has weekly IRC meetings in alternating time zones.

Members

No formal membership is required. Please introduce yourself and/or fellow colleagues to this working group using one of the mailing lists below, or by attending one of the IRC meetings.

This group is open to all members of the scientific OpenStack community and supporting vendors.

Documents

The Scientific Working Group maintains a guide to meeting the requirements of scientific computing workloads on OpenStack, titled The Crossroads of Cloud and HPC: OpenStack for Scientific Research.

Contents:

• OpenStack and Virtualised HPC
  • The Overhead of Virtualisation
  • Using OpenStack to Deliver Virtualised HPC
  • Virtualised HPC on OpenStack at Monash University
  • Optimising for “Time to Paper” using HPC on OpenStack
  • Further Reading
  • Acknowledgements
• OpenStack and HPC Network Fabrics
  • Using SR-IOV for Virtualised HPC Networking
  • Virtualisation-aware MPI for Tightly-Coupled Cloud Workloads
  • Infiniband and other Non-Ethernet Fabrics
  • An RDMA-Centric Bioinformatics Cloud at Cambridge University
  • The Forces Driving HPC and Cloud Diverge in Network Management
  • Further Reading
  • Acknowledgements
• OpenStack and HPC Infrastructure Management
  • Deploying HPC Infrastructure at Scale
  • Bare Metal Management Using OpenStack Ironic
  • Building Upon Ironic to Convert Infrastructure into HPC Platforms
  • Most of the Benefits of Software-Defined Infrastructure...
  • Further Reading
  • Acknowledgements
• OpenStack and HPC Workload Management
  • Workload Management in HPC and Cloud
  • Delivering an HPC Platform upon OpenStack Infrastructure
  • Cloud Infrastructure Does Not (yet) Provide All the Answers
  • Summary
  • Further Reading
  • Acknowledgements
• OpenStack and High Performance Data
  • File-based Data: HPC Parallel Filesystems in OpenStack
  • Applying HPC Technologies to Enhance Data IO
  • Optimising Ceph Storage for Data-Intensive Workloads
  • The Cancer Genome Collaboratory: Large-scale Genomics on OpenStack
  • CLIMB: OpenStack, Parallel Filesystems and Microbial Bioinformatics
  • Further Reading
  • Acknowledgements
• OpenStack and Federated Identity Management
  • Federation Terminology
  • The Concepts of Identity Federation
  • Federated Identity Management in OpenStack
  • Case Study 1
  • Case Study 2
  • Further Reading
  • Acknowledgements