Oxford University

The researchers from the Department of Materials collect the data from their experiments at facilities in the UK and continental Europe, and analyse them in collaboration with different research groups across the university. Transferring data between different sites and buildings by portable hard drives was time-consuming, prone to mishandling and could be difficult to keep track of their locations. "A colleague from University of Southampton, who faced similar difficulties, suggested the Synology NAS server, which had helped them solve such problems," says Mostafavi. "After consulting with the IT services in the Oxford Materials department, the decision was made to purchase a DS3612xs unit.

After installing the Synology unit, Mostafavi was immediately able to transfer 7 terabytes of data within a week. He then had four high performance computer workstations connected to DS3612xs, analyzing the data in different ways. "Some of these analyses are computationally expensive and require much RAM, typically more than 48 GB, so the computers involved in the analysis can't be used for any other task," says Mostafavi. "Also, if the data were on an external hard drive attached to the computer, access to it by other users would be virtually impossible as access speeds become very slow and would essentially develop a bottleneck that slows down the studies significantly." As a high performance storage server that is designed to handle demanding tasks autonomously, the DS3612xs is able to meet with the department's high standards without affecting the workstations' performance during data transfer.

Mostafavi adds: "Now at any given time, two to three researchers can be working on different parts of the data, conducting their own analyses in parallel. Shared access therefore increases the speed with which the whole process is completed. It gives the researchers more flexibility for managing their time as they can access the data whenever they decide."

The research group is expanding in members and collaborations. "More experiments are being performed, and data sets are increasing in size. It is very likely that we will expand the unit so the parallelization of work can continue," says Mostafavi.