Research Infrastructure

Dedicated Research VM Infrastructure

This article describes the existing Research IT Virtual Machine service and outlines our planned developments. For more information, please email


its-ri-team@manchester.ac.uk

Existing Research Virtual Machine (VM) Service (RVMS)

IT Services already offer a research VM service. Through this service, VMs are offered to research groups:

  • with a relatively low specification, such as one or two CPU cores, four GB RAM and 25 GB local storage;
  • on which researchers may have admin/root privileges.

A standard Scientific Linux (cf. RedHat Enterprise) image and MS Windows are offered pre-installed on VMs – other Linux distributions may also be installed by IT Services on a best-effort basis (OS-X cannot be installed for licensing reasons).

Uses to which this infrastructure can be put include:

  • Web servers which cannot be accommodated on the shared hosting service [[…URS???…]], e.g., those on which researchers require administrator/root privileges.
  • Shared development areas for research groups and their (external) collaborators.

Planned Developments

The existing service is limited in terms of the specification of VMs which can be offered, in particular, CPU-intensive work cannot be carried out and high-memory and/or storage requirements (e.g., for large databases) cannot be accommodated. Therefore, in order to address a wider range of requirements, we plan to run a contribution-based Research VM Service based on dedicated Research VM Infrastruction (RVMI), using a financial model similar to that of the Computational Shared Facility.

The RVMI:

  • Will be run on a sustainable, contribution-based model: research groups contribute funds which is pooled to procure hardware; groups then receive VMs equivalent to the funds contributed; seed money has already been obtained from The University (from a successful case made to ISSC).
  • We expect to be able to accommodate VMs with, for example, one TB of local storage and 128 GB RAM (e.g., for a Web-fronted SQL database).
  • CPU-intensive work will also be accommodated
  • We also plan to offer job-submission gateways to computational facilities such as the CSF and Hydra (e.g., Web-based and/or drag-n-drop) which should make use of these facilities easier for novice users (albeit for a limited range of applications and job types).
  • Finally, we will explore the possibility of offering a visualization service through this infrastructure.

Comments are closed.