Fixed-content data archiving products get a boost

Caringo, EMC and HDS spruce up their fixed-content data archiving products as buzz around compliance and legal discovery continues.

Storage vendors are polishing up fixed-content data archiving products this quarter to be better prepared to ride the recent wave of discussion around data archiving for compliance and legal discovery.

The most recent product update, announced today, came from content-addressed storage (CAS) upstart Caringo Inc. Caringo was founded by Paul Carpentier, who sold another fixed-content archiving product called FilePool to EMC Corp. in April 2001 for a little under $50 million. FilePool became EMC's Centera CAS product.

Carpentier's new company first announced its CAStor product last fall. Now, according to Mark Goros, CEO and co-founder of the company, the updates in version 2.0 of CAStor are intended to make CAStor "a more world-class system." These updates include local and wide area replication, the ability to do bare-metal restore for databases within the system, and scalability and management brush ups.

More on fixed-content archiving
New privacy rules may complicate records management

EMC reports Clariion surge, data archiving slump

EMC floats raft of updates for disk arrays

Google buy shakes up email archiving
A software-only product, CAStor 1.0 came on a memory stick that could be plugged into a wide variety of off-the-shelf computers. But according to Goros, the company found itself running up against scalability limitations of its memory stick schtick in the months since the first release. The installation of the software from a USB stick in under 60 seconds looked good in demonstrations but wasn't quite as attractive when it came to installing bigger systems. CAStor 2.0 still comes on a memory stick, but the product can now be installed on multiple nodes from a central management console.

Goros said previously the company had felt there wasn't a need for a management console, since the product doesn't require application vendors to write to a proprietary API to access the system. That idea was also soon shot down by end users. "Before, you had to ping each node for a status report to monitor and manage the system," he said. "We didn't think it was a lot to manage, but these are things people like to track centrally."

Also, Goros admitted the ramp up for the company has been a frustrating process. Caringo is sticking with a channel-only model, banking on selling the product through third-party independent software vendors (ISV) who may build packages around CAStor and hoping that will eventually lead to wider distribution. Currently, the company claims 35 end users, few have been publicly announced.

"I almost wish sometimes we'd changed the strategy and sold to end users first," Goros said. "It's frustrating in the short term, but building applications based on CAStor means many more end users out there in the long term."

While Caringo has some advantages, including vastly lower cost than Centera and similar hardware-based CAS systems, it still has quite a way to go, analysts said.

"They might have a few customers. However, I would be surprised if their traction has actually picked up significantly since their launch a year ago," said Greg Schulz, founder and analyst with the StorageIO Group. "[This announcement] is keeping up with the Joneses, but what they need to do is add customers and real OEMs."

Centera: The CAS front runner, but growing older

Other experts believe there may be an opening for a new product in the CAS space, yet. EMC surprised many industry watchers with its Centera announcement last week, which was almost entirely hardware-focused, adding denser, more power-efficient nodes (called Generation 4 Low-Power nodes) to the system.

It had been anticipated that Centera's MD5-based hashing algorithms would be brushed up and that software changes would be made to improve the system's performance, according to John Webster, principal IT advisor with Illuminata Inc.

"I expected more software development with the latest release of Centera," Webster said. "In the end, it wasn't a very significant refresh."

EMC's recently announced earnings numbers also show that its archiving business over the last quarter has not been stellar. Compared with 20% overall revenue growth, as well as 34% growth in midrange storage area network (SAN) data storage systems, the content management and archiving business unit at EMC grew just 5%, prompting CEO Joe Tucci on last week's earnings call to declare that the division "can and will do better in this space."

HDS: Building up a partner program

Hitachi Data Systems (HDS), a much larger competitor to EMC than Caringo, has also been trying to strengthen its challenge to Centera in recent weeks. In May, HDS unveiled version 2.0 of its Hitachi Content Archive Platform (HCAP), based on its acquisition of Archivas Inc. in February for approximately $120 million. Since then, more ISVs have been porting applications to the product, including Bus-Tech Inc., Princeton Softech Inc., National Digital Medical Archive Inc. (NDMA) and Eastman Kodak Co. HCAP has also added support for more products from existing partners, including Symantec Corp.'s Enterprise Vault.

Though the number has doubled in the last quarter, HDS' current count of 50 partners still doesn't match the hundreds that have ported applications to Centera, but the company argues that since HCAP operates as a standard filesystem, users can manually migrate files to it without the need for an API between vendors. The ISV partner program is in place so that application vendors' products, such as Symantec Enterprise Vault or Mimosa's NearPoint, can automatically control the migration of files based on policy, according to Charles Hickman, HDS' senior director of business development.

Like Centera, HCAP uses a hashing algorithm to encrypt and single instance data, but it does not perform content addressing the same way Centera and CAStor do. HCAP retains original file names rather than replacing them with a numerical identification and uses a WORM filesystem to fix data in place in the archive and render it immutable.

"I would compare HCAP more to EMC's Infoscape product than Centera," Webster said. "Centera is a single-platform device, while HCAP is more of an 'umbrella' multiplatform application."

Generally, Webster said, users should take some of the hype around archiving products with a grain of salt. As with any trendy topic, "there are vendors out there who want to create a cloud around this subject so they can sell products to dispel it," he said.

However, he said, "It is critical that users get a handle on a long-term data archiving situation that in many cases is still a pile of backup tapes. That's just not rational anymore."

Other products in the fixed-content archiving space include Hewlett-Packard Co.'s (HP) Reference Information Storage System (RISS), IBM's DR550 and Nexsan Technologies Inc.'s Assureon.

Read more on Integration software and middleware