Dedupe start-up GreenBytes gets green light

Only iSCSI de-duper in the world

Start-up GreenBytes has announced its GB-X products that deduplicate primary data, both NAS and iSCSI SAN, and which are based on an enhanced ZFS file system.

GreenBytes was founded in 2007 by a Robert Petrocelli. He previously founded Heartlab, a Rhode Island-based health-care IT business. This was sold for $132.5m in 2005 to Agfa-Gevaert. GreenBytes is his next enterprise and involves a storage server, originally using Sun's X4500 Thumper box and Open Solaris. It has now switched to a vanilla x86 platform from an unidentified source, following uncertainty over Sun's hardware product supply due to the Oracle takeover.

There are two models: the GB-2000 and GB-4000. Both run the GreenBytes File System (GBFS), an enhanced ZFS (Sun's Zettabyte File System), which provides inline deduplication of incoming primary data or backup data. It also provides power management that enables a subset of the drives to run in a low-power mode, making the products greener.

The software protects the system against double drive failure and there's no single unprotected point of failure. ZFS' block hashing helps data integrity, as does proactive disk scrubbing.

The GB-2000 is Xeon 5520 processor-based, with what GreenBytes describes as eight virtual cores. It ingests data at up to 650MB/sec, using an optional 10GbitE interface. The larger GB-4000 doubles up on the Xeons to 16 virtual cores, with an ingest rate of 950MB/sec.

The 2U GB-2000 has up to 24 2.5-inch drives, an SAS backplane and five 1GbitE ports. It can scale from 12 to 60TB of capacity. The 4U GB-4000 can scale from 24 to 216TB of capacity, supporting up to 48 hard drives and two 10GBitE ports, as well as the five 1GbitE ones. It can also support an optional solid state drive cache of undisclosed capacity. Its SAS backplane runs at 24Gbit/s, double that of the GB-2000.

Both systems are managed from Microsoft's Management Console via plug-ins. All software features are bundled in.

GreenBytes also announced the completion of an $8m A-Series funding round, with $7.5m of the cash having been raised in May. That means initial or seed funding came from elsewhere, possibly using the cash from the Heartlab sale. The product technology is largely software-based with an optimised commodity hardware base.

As deduplication systems, the GB-X products will nominally compete with inline de-duping Data Domain and Quantum, primary data de-duping NetApp, and then other deduplication suppliers. We suggest you view them possibly as serious server/storage engines for data rich applications. How much processing horsepower is left over from deduplication/rehydration to run applications of any substance will need to be tested by potential purchasers.

They look to be products for storage system integrators and VARs to use as an application base, with the pitch being far lower power consumption than competing arrays and more efficient use of storage capacity.

An entry-level GB-2000 is priced from $65,000 with the GB-4000 starting at around $100,000. Even though neither is cheap, they will probably turn out to represent better value for money on acquisition price and ownership cost bases than existing mainstream storage array vendors' products. However, you have to bite the bullet of an unknown and untested vendor.

It looks like they will mainly be the territory of skilled VAR and system integrators, or similarly skilled end-users, for now. ®

Other stories you might like

Biting the hand that feeds IT © 1998–2021