Currently, IT managers have to back up their data, then add more space, then restore the data across the new drives. This can take hours and hours if theres terabytes of material. There are even more laborious ways to do backups, too, which is why OSD is so compelling.
Essentially, OSD removes the file system management component (the piece that holds the "meta data" information about how important data is and where it belongs) from the operating system and moves it onto the storage device. This means the storage device is now intelligent enough to manage the data itself rather than relying on the operating system to do it.
Panasas says this will lead to marked improvements in manageability, performance, and compatibility of storage networks.
With this model in mind, Panasas is building two hardware devices: a Smart Drive that will hold multiple gigabytes of data; and a Storage Manager, or virtualization engine, that will make all the Smart Drives on the network appear as one pool of storage to multiple servers accessing that storage. Distributed file system software, which is 80 percent of the solution, the company says, makes it possible to drop in more Smart Drives as needed. These are then seamlessly assimilated into the network.
Panasas and Carnegie Mellon University are not the only ones working to define the technical specifications for OSD. Hewlett-Packard Co. (NYSE: HWP), IBM Corp. (NYSE: IBM), Quantum Corp. (NYSE: DSS), Seagate Technology Inc., and StorageTek (NYSE: STK) are also pushing the standard. The American National Standards Institute (ANSI) T10 committee has been set up to approve the spec, which its backers hope will happen by the end of this year.