EMC is expected to announce as early as next month an upgraded version of its fixed-content storage array that boosts its scalability, availability and reliability, and preserves the integrity of stored data.
While EMC declined to comment on Centera V2, sources say the product will add these hardware and software features:
- Significantly increases the capacity of Centera V1 for storing even larger amounts of fixed-content data - data that does not change - such as CAT scans, seismic data or video files.
- An AC transfer switch, which fails over the array to a secondary power source if the primary power fails.
- SNMP support to notify the systems administrator of problems that are occurring or that the Centera is automatically fixing.
- Software features that protect data if a segment of storage fails, reclaim disk space when data is no longer needed, shred data so it can't be recovered and improve replication capability to eliminate interference with operations.
Centera and other storage devices from vendors such as Network Appliance Inc. and StorageTek Inc. use inexpensive Advanced Technology Attachment drives to store data online and close to its source, rather than cumbersome and slow-to-retrieve tape or expensive primary Fibre Channel or SCSI storage. By year-end, fixed content, also called content-addressable storage, will comprise as much as half of all data worldwide, according to a University of California, Berkeley study.
Paul Galjan, an IS director for Science Applications International Corp., on contract to building materials manufacturer USG in Arlington, Va., has looked at EMC's Centera to manage his fixed-content application. He ultimately chose an LSI Logic Inc. array to store his fixed content because the ATA drives did not give him the scalability he needed a year ago when he bought the LSI box. He expects that has been fixed now.
"Those users who are interested in Centera are [often] interested in capacity and price," Galjan says. "[Increasing] the capacity of the drives will reduce the number of components, thus reducing price."
The new Centera, which contains an upgraded Pentium III processor running at 1 GHz, increases disk capacity as much as 56 percent by using 250G-byte drives. A fully configured Centera now can support 32 terabytes of disk capacity.
Among the most important features of Centera V2, users say, are the software enhancements. EMC has added content-parity protection, in which a data object is written to disk as six fragments and a parity fragment. Parity is a technique for checking that the data written to disk matches the data sent to disk. With content-parity protection, if a segment of the disk fails, the original data can be reconstructed from the fragments and the parity information. The new Centera also integrates a garbage collection feature - when Centera detects that data is no longer used or referenceable, it will collect that data and dump it, effectively recovering disk space.
Centera uses a digital-signature algorithm called Message Digest 5 (MD5) to identify data objects on disk. Applications writing fixed content to Centera use MD5 to access and retrieve data objects. Because each MD5 address is intended to be unique, EMC has added the ability to avoid collisions and data loss that could occur if the application generated data objects with the same MD5 signature.
Pricing for the upgraded Centera was unavailable. The current version, Centera V1, starts at US$205,000.