By Ann Silverthorn
—Kazeon today introduced SnapSearch and Recovery software, which classifies and indexes backup and archive data and supports Network Appliance's data-protection products, including SnapShot, SnapVault, and SnapMirror. Kazeon's announcement builds upon its October 2005 agreement with Network Appliance, in which NetApp agreed to OEM Kazeon's Information Server (IS) 1200, an out-of-band storage appliance.
"Many companies have deployed disk-based backup solutions, and data is often dispersed in multiple remote offices that are also backing up a smaller set of online data," says Michael Marchi, vice president of solution marketing at Kazeon. "When a centralized help desk receives a request to recover a file, it's a manual chore. Even if the information is online, administrators have often had to remotely mount a file system and traverse a network to find the information. Help desks need to provide faster recovery for their end users."
SnapSearch and Recovery enables users to search all online backup data. It classifies and indexes information so it can be recovered rapidly, regardless of where it's stored on the network. The software searches both the metadata and the contents of archived e-mails, Word documents, PDF files, and others types of files. Although some companies may keep many versions of a file, SnapSearch and Recovery identifies the unique versions. To recover files, users double-click the files to open them, or they can right-click the files to save them. The software allows end users to recover their own files without the assistance of the help desk, or, if more control is desired the company can give this permission to the help desk only.
Laura DuBois, research director of storage software at International Data Corp. (IDC), says companies will face data growth rates of approximately 52% over the next 12 months and have a limited number of people to manage the capacity. "Larger firms may have dozens different NAS devices containing unstructured data. They have a good understanding of their structured data in databases, but need a better understanding of their file-based unstructured data to address the business risks associated with regulatory compliance, electronic discovery, and potential leaks of sensitive corporate data.
"That's where Kazeon's and other content-based, index approaches come into play," says DuBois. "There's a clear line of demarcation—some products classify data based on the traditional file system metadata, such as file name, age, size, etc. That's useful, but it doesn't help with the business issues of unstructured data. SnapSearch and Recovery cracks open files, understands the content in the files, and flags them based on classification schemes. It can help point out where the risks are within huge volumes of unstructured data."
DuBois says that Scentric and Avalere are two companies in stealth mode that she thinks will be introducing content-based classification products in the near future, although she notes that Kazeon has a large time-to-market lead. She also cites data classification vendors such as Abrevity and UK-based Njini as players in this market.
Comparing EMC's eDiscovery to Kazeon's SnapSearch and Recovery, DuBois says that "eDiscovery focuses mainly on e-mail, because that's the number-one target for litigation. eDiscovery builds a legal-matter repository on EMC storage that allows for easy discovery, audit trails, and maintaining the chain of custody for everything relative to a particular legal matter. Kazeon's SnapSearch and Recovery discovers and builds an index and classification scheme around the data, but doesn't build a separate repository." (For more on EMC's eDiscovery, see EMC adds eDiscovery services.
A short list of other vendors of data classification tools includes Arkivio, CommVault, EMC, Index Engines, and StoredIQ.
Kazeon's SnapSearch and Recovery, including the Information Server, is priced at $60,000 per node. The number of nodes required depends on the number of files, and the system supports 30 to 80 million files, depending on the file type. Kazeon's Marchi says that most customers buy three or four appliances per data center.