Mission + Future
The DORS project aims to provide centralized, professionally managed, enterprise-grade high performance data storage at a cost that is less than or equal to purchasing and maintaining local, entry-level storage, but without the added complexity and headaches of doing so. The service is designed to:
- Provide high-speed campus-wide accessibility to/from ACCRE, data collection/core facilities, and end-user workstations.
- Have no single point of failure, including but not limited to file servers, storage controllers and network switches
- Support nightly backups of all data, at a different physical location from the primary disk.
- Scale to multiple petabytes of capacity, with a self-sustaining financial model.
This project is a collaboration between the Center for Structural Biology (CSB), the Advanced Computing Center for Research and Education (ACCRE), and Vanderbilt University Information Technology (VUIT). It was funded by NIH S10 RR031634.
GPFS and TSM together support “Information Lifecycle Management” (ILM) technology. ILM enables data on disk to be marked for long-term offline archival to tape. In this scenario, the GPFS filesystem leaves behind a “stub” that appears as a regular file to the user. By interacting with this special stub file, users can gain unattended access to archived files after a delay while TSM automatically retrieves the data from tape and places it back on disk. This feature has not yet been enabled or tested by the DORS team, but it is a priority item that we plan to explore over the next several months. If successful, this could change the way we all manage our data by automating and thus dramatically reducing labor associated with maintaining long-term data archives. This could have a positive impact on maintaining compliance with funding agency regulations around data retention.
The DDN SFA-12K40 is expandable to 20 84-drive bays for a total of 1680 disk drive slots. If the system is successful over the next few years with demand beyond the existing 840 slots, this is an upgrade that that could bring the total disk capacity to in the neighborhood of 2PB or more, depending on the individual disk drive speeds and capacities that will be available in the future.