Nexsan has introduced a storage subsystem which it said could enable users to consolidate their long-term business-critical data stores into a single consolidated platform.

The device, called DATAbeast, couples a virtualizing storage controller that's capable of thin provisioning, storage pooling, mirroring, snapshots and replication, with Nexsan's SATA and SAS-based AutoMAID disk arrays. Servers can connect to it via either 4Gbit/s Fiber Channel or Ethernet.

The result, claimed Nexsan's Andy Hill, is low power, high density storage, but with the sort of centralized management tools currently only found in high-end storage subsystems such as IBM's ESS (Shark), EMC's Clariion and the .

DATAbeast is aimed at small and mid-sized organizations which currently use the likes of SATA disk arrays or NAS filers to store bulk data - data that once it's created, changes infrequently. According to Nexsan, examples of this include business records, medical and scientific data, presentations, and audio and video archives.

Some of these organizations now have so storage boxes that management is becoming a problem, while others are realizing that consolidating both transactional data and near-line data onto the same high-performance filer doesn't make economic sense, Hill said.

"People have now moved from consolidating everything onto one box to, well, one size doesn't fit all," he added. "DATAbeast is Tier 2 storage with Tier 1 features. It's an appliance-type solution, like a SAN in a box with all the levels of fault-tolerance you need, snapshots, virtual storage pools and so on.

"However, Nexsan's strategy is still Tier 2, sitting next to the Tier 1 players. Tier 1 is all about performance - so it's Fiber Channel drives, SRDF replication, high levels of consultancy, etc."

Hill said that as well as database or archive storage, DATAbeast could also provide the back-end storage for a de-duplication server or a virtual tape library (VTL).

He added that DATAbeast costs between US$1800 and $2700 per TB, depending on its capacity and the type of disk arrays used. The starting point is a subsystem holding 42TB, but it can be expanded to 336TB in a single 42U rack, and racks can be clustered to accommodate up to 4PB in a single virtual system.