Mission

To provide flexible high performance computing power for computational research at Harvard Medical School in a cost effective and collaborative manner

Summary

The Orchestra shared research cluster is a service provided by the Research Information Technology Group (RITG) at HMS. Utilizing the Linux operating system and LSF queuing and cluster software from Platform Computing, Orchestra provides an open and powerful environment for researchers to conduct computational research. Researchers may choose from a library of common biocomputing applications that are pre-configured on the cluster, can recommend additional applications to be included for all researchers, and in some cases arrangements can be made to install commercial software with special licensing arrangements for a lab or group. The cluster will also communicate with other clusters at HMS and elsewhere on an experimental basis.

For more detailed information about Orchestra's technical specifications and available applications, please visit the Orchestra wiki: http://wiki.med.harvard.edu/Orchestra

• eCommons login credentials are required for access to the Orchestra wiki
• Most researchers at HMS-affiliated institutions can self-register for an eCommons account. Please visit the eCommons website for more information.
• If you do not have eCommons credentials, but would like more information about Orchestra, please contact RITG via this support form

Please use this form to request an account on the Orchestra cluster.

Service Model

The cluster is available to any HMS affiliated researcher. Researchers will be able to use all available computing power – but jobs will be balanced when all capacity is in use. Researchers may also elect to purchase “dedicated” nodes for which they will always have priority access – though they will also be available for other users of the cluster when otherwise idle. Conversely, a job run by a researcher with dedicated nodes will take advantage of non-dedicated idle nodes.

Usage Costs

Currently, system accounts and all related resources (e.g. storage, backups, network, technical support, data center hosting) are provided free of charge for all users, with one exception: Fee-for-service charges do apply to those users who require especially large storage resources from the cluster. While HMS does reserve the right in the future to impose other fee-for-service charges as needed, it is our intention to keep services available for free as much as possible.

• Technical support is available through the RITG, and all requests for support or information should be initiated via this support form

Project Justification

Computing infrastructure is no longer just a way of supporting research – but it has becoming a critical tool for conducting research. With the advent of computational methods, the use of high performance computing methods is essential to maintaining Harvard Medical School at the forefront of bioinformatics and computational research.  While some labs have the resources to set-up and maintain their own computational cluster infrastructure, smaller labs do not. Moreover, there are economies of scale to be gained by centralizing these efforts – thereby eliminating redundant systems support personnel, redundant server rooms, and unnecessary power and electrical burdens. By setting up a new research cluster, and by joining it with existing and planned clusters on the Quad, researchers will be able to take advantage of spare CPU cycles while guaranteeing access to the CPUs that they have contributed themselves. In turn research will be conducted faster and the results will be available for publication sooner. This cluster represents the commitment of HMS administration to provide the tools and services needed by the HMS faculty to continue to conduct ground breaking and high quality research.