Google Announces Launch of New Search Appliance

10 million documents in a single box.

by / August 7, 2008 0

Photo: Google Search Appliance hardware

Google Inc. announced today the launch of the latest generation of the Google Search Appliance, an all-in-one enterprise search solution that can index up to 10 million documents in a single device. This iteration of the Google Search Appliance bests its predecessor's document capacity by 7 million.

As enterprise content continues to be created, more repositories are put online to manage the torrent of data. Typically this means organizations must build ever more elaborate infrastructure. In some cases, enterprise content management can result in the construction of something like a dedicated data center. For the end-user within an organization, simply searching for a specific document can become a headache.

Nitin Mangtani, Google's lead product manager of enterprise searchPhoto: Nitin Mangtani, Google's lead product manager of enterprise search

"We are now able to scale our single-node appliances to 10 million documents," said Nitin Mangtani, lead product manager of Enterprise Search at Google. "The old boxes ... could go to 3 million. Now we have come out with brand new hardware and software architecture allowing us to scale to 10 million documents in one single box." Mangtani added that Google is offering clustered solutions capable of managing up to 30 million documents or more.

In tandem with the Google Search Appliance, the company also announced the launch of new, personalized search capabilities available within an organization running an appliance. Users can specify -- or administrators can specify for them -- what sort of search results should be considered most relevant.
"If somebody in engineering is doing a search, he would see design documents shown higher versus somebody in marketing or sales. In those cases, the marketing documents and cases studies will show higher," explained Mangtani.

In addition to making enterprisewide document searches customizable, the Google Search Appliance software includes advanced biasing controls. Metadata values can be changed on the fly. For example, if an administrator wants search results to boost documents written by a specific employee, all he or she needs to do is adjust a values slider and select the employee name.

But with only one box, what about disaster recovery? Few people would be ready to load 10 million company documents onto the appliance without some reasonable assurance a broken water pipe won't wipe out all records. According to Google, each Search Appliance sold will include a second appliance unit, which can be kept in a separate data center.

Google said the boxes also feature a host of security features, such as Kerberos, lightweight directory access protocol (LDAP), BASIC programming language and NT LAN manager (NTLM) authentication, PKI authentication with x509 certificates, and Windows Integrated Authentication. Document-level security makes it possible for administrators to, for example, easily ensure employees can only search certain documents. Sensitive data can be made virtually invisible to those who aren't authorized to see it.

"Let's say I was trying to do a search on a defense contract," Mangtani hypothesized. "If I was looking for information I didn't have access to, I won't even know those documents existed."

Mangtani said the Google Search Appliance requires only modest technical expertise and can be set up in less than a day. The devices start at around $30,000 and include two years of support.

Chad Vander Veen

Chad Vander Veen previously served as the editor of FutureStructure, and the associate editor of Government Technology and Public CIO magazines.