Bright Computing, a provider of cluster and cloud infrastructure automation software, announced that Data in Science Technologies has implemented Bright Cluster Manager at the HudsonAlpha Institute for Biotechnology.
The HudsonAlpha Institute for Biotechnology is a non-profit genomics and genetics research institute located in Huntsville, Ala. Its mission is four-fold: genomic research to improve human health and wellbeing, genomic medicine, economic development, and education.
Data In Science Technologies (DST) has been providing HPC compute and GPFS cluster support to HudsonAlpha for two years, to help the biotech institute maximise cluster availability, performance, and configuration consistency. When HudsonAlpha set DST the challenge of further improving cluster efficiency and scalability, DST recommended Bright Cluster Manager.
Katreena Mullican, Technology Director at HudsonAlpha, explained the Institute’s requirements. ‘During the proof of concept, we paid close attention to performance, reliability, ease of configuration from both graphical interface and command line, ease of troubleshooting, and integration into our existing environment. Bright met all these requirements.’
Bright Cluster Manager also fulfilled some specific needs; namely to run in a virtualised environment and to provision HudsonAlpha’s customised operating system (OS) onto a variety of compute cluster hardware. Additionally, HudsonAlpha aims to treat infrastructure as code, so the biotech institute prefers tools that offer a command-line interface allowing processes to be scripted.
DST carried out the implementation project for HudsonAlpha and utilises Bright to manage the day-to-day operations as part of its HOMS (Operational Management System) offering.
Previous provisioning methods at HudsonAlpha included adapting existing automated OS installation workflows for cluster use. This provided OS installation, initial package installation, and initial network configuration. Once installed, separate applications were used to identify and track individual cluster nodes. Ongoing maintenance, including the rollout of new packages and networking changes, was accomplished by another application. Provisioning a single cluster node and integrating into all the existing applications took hours per node. Under Bright Cluster Manager, deployment of a new cluster node is complete in less than 10 minutes. Maintenance activities, such as adding packages or making network changes, is accomplished cluster-wide within minutes.
Don Schrenk, Senior Partner at DST, commented: ‘Bright enables the HudsonAlpha team to spend less time focusing on cluster configuration and more time working directly with researchers, while DST ensures that the HudsonAlpha cluster design meets the needs of the Institute. We are delighted that our partnership with Bright is providing such a significant value to another of our customers.’
Clemens Engler, Director Alliances at Bright Computing, added, ‘HudsonAlpha places high demands on its compute infrastructure. We are pleased that DST has introduced and leverages the Bright technology to enable the Institute to meet its computing needs.’
