In Search of... Enterprise Search

In a spate of recent announcements, storage networks are taking center stage in IT's search for a better way to, well... search enterprise data.

Startup Index Engines Inc. is the latest of several companies to unveil products that use SAN or NAS as the starting point for indexing corporate data and making it searchable (see Index Engines Unveils Search Architecture). The idea is that if documents and email can be identified by keywords and other criteria, they can be more easily managed as part of a larger tiered storage or compliance strategy.

Index Engines is basically expanding on an earlier announcement (see Index Engines), providing details about its appliance, which sits between a server and a device like a tape library or disk array, scrutinizing unstructured data as it is backed up on a Fibre Channel link.

Being in-line with backup has a range of advantages, according to cofounders Tim Williams and Gordon Harris, whose past projects included launching Tacit Networks Inc. and CrosStor, which was sold to EMC Corp. (NYSE: EMC). Index Engines can build its own metadata from the info gleaned from the backup traffic. It can adopt authentication in use on the SAN. And it can locate the searchworthy data without going to a bunch of other devices and hosts to get it.

CEO Williams concedes that putting his appliance in front of backup adds "several minutes per hour" to the backup process. But he also says the unit, priced from $29,500, can handle 500 queries per second. In contrast, the Google Search Appliance, a LAN-resident unit from Google (Nasdaq: GOOG), handles 1,000 queries per minute, according to Google's Website.Index Engines isn't alone in seeing a customer need for storage-oriented enterprise search engines. Yesterday, Kazeon Inc. emerged from stealth with claims of a pending solution (see Kazeon Comes Out). And earlier this year, StoredIQ announced plans for an appliance of its own, geared to specific vertical markets and bulked up with security (see Deepfile Becomes StoredIQ).

There are other solutions, too, such as EMC's recently announced Centera Seek program (see EMC 'Charges' Into Archives), which could point to a future trend to link enterprise search with specific backup, archiving, or CAS (content addressed storage) wares. Indeed, Williams says the partnership opportunities in this space are very good and that his firm is talking to at least one potential candidate.

The growing niche for enterprise data search engines based on storage competes with a range of higher-end solutions that aren't storage-oriented. Software from the likes of Verity Inc. and FAST, for example, deals not only with unstructured but structured data and includes custom services, hosting, and a range of other solutions geared to making searching enterprise data a strategic solution, often tagged to a specific market.

So far, it doesn't look as if the storage-based products can offer the range of solutions these higher-end companies can provide. The segment now sports a smattering of products with particular angles on the problem of making enterprise data searchable. But until there is a better range of complete choices -- ones that include security, automated policy enforcement, and other features -- storage managers will be faced with an uneven set of choices.

Mary Jander, Site Editor, Byte and Switch0