05 November 2019

Imperial College London’s Research Computing Service (RCS) – part of the ICT department – plays a vital role in addressing the computing and storage needs of the research community. 

Its academic research community had been served by a centralised compute service under the management of the Research Computing Service, with users charged based on reserved capacity.

The system gradually expanded over time to address incremental growth and any storage attached to it only served the purposes of transient data storage.

However, it resulted in a very complicated and fragmented environment, with over 30 separate independently managed islands of storage, difficult to access, manage and use as well as expensive to run.

Faced with poor performance and the high-costs of data-centre space, academics had storage capacity issues. 

Indeed, many users treated the compute service as a de facto storage repository for all their research data and many petabytes of data had been built-up over the years with neither structure or process.

As a result, it was becoming difficult for the RCS team to gauge whether data was in active use, had been abandoned years ago, or whether it was hot or cold.

The idea was that by moving to a new single, centrally-managed and supported system, the university would ensure that it kept happy the data providers and funders who expect researchers to demonstrate responsible data management, as well as complying with increasingly stringent regulations for the responsible handling of personally identifiable information (PII).

The Research Data Store (RDS) was designed to address these challenges and provide a longer-term strategy to manage and store research data.

Furthermore, the goal was to enable researchers to access data with ease and speed and store that data throughout its life-cycle whilst enabling the RCS to intelligently manage growing storage demands and efficiently recover costs.

ArcaStream – supported by Imperial’s integration partner Tectrade - was selected to provide a scalable research storage solution to seamlessly integrate legacy infrastructure and support the university’s future storage strategies.  

PixStorTM, ArcaStream’s scalable storage platform based on IBM Spectrum ScaleTM parallel file system, combines flash, disk, tape and cloud storage into a single global name space.

With a software- defined architecture, it uses open standard commodity hardware to avoid vendor lock-in coupled with powerful data management tools – including tiering, cloud integration, monitoring, search and analytics – to drive workflow efficiencies and reduce costs.

Now, with PixStorTM, Imperial can scale to meet future requirements with confidence.