A Gateway to Unstructured Data
The ability to access the recent explosion of unstructured data, in the forms of social media data as well as video and audio files, can be limited. SGI, with their planned release of the InfinteStorage Gateway on June 15, hopes to quicken that access, with the goal being to provide a single universal place from which to draw all sorts of media types and files.
SGI built the Gateway by introducing software to its Xeon-powered 4U MIS Storage Server, which is capable of storing 276 terabytes in disks and solid-state drives. According to SGI, the Gateway intends to sell to those with “multi petabyte infrastructures spanning media, life sciences, manufacturing and other data-intensive industries.”
Further, the InfiniteStorage Gateway is slated to support files from many sources, including the (relatively) high latency, low speed sources of cloud and tape storage as well as high-speed object and MAID (Massive Array of Idle Disks) storage.
Of particular interest is the media aspect. The ease with which video and audio files can be created and uploaded contributes significantly to the big data explosion of the last few years. A lot of that media ends up in higher latency tape and cloud storage facilities.
However, life science research that work with SGI would look to employ the Gateway to gain faster access to that data, and things like playback of dynamic cell and DNA structures. That can be facilitated by the IT managers running instances of the Gateway, who can place restrictions and policies upon the data retrieval based on relevance. This is done through SGI’s relationship with Scality, which allows the aforementioned object storage. The API interface supports Scality’s RING object storage, which is reportedly highly scalable.
Further, SGI plans to integrate other object storage interfaces such as S3, CDMI, and OpenStack in the near future.
This SGI Gateway is competing with Quantum’s StorNext platform in the attempt to virtualize file storage in a big data world that may find such virtualization increasingly useful. Providing, well, gateways between data sources and heterogeneous file types would ease many institutions’ transition into storing and accessing all of their big data.