Google Revs Up Searches

Googles new search appliance provides businesses with a very powerful but simple-to-deploy system for indexing and searching large and complex corporate Web sites, portals and intranets.

In tests of the 1U (1.75-inch) GB-1001 rack-mounted appliance from Google Inc., eWEEK Labs was impressed with its ability to quickly crawl and index very large Web sites. We also liked that our results benefited from all the features one expects from a Google search, such as cached pages and HTML conversions.

However, although the appliance will meet or exceed most companies requirements for a search engine, it lacks some advanced features found in other enterprise search engines, including the ability to dynamically create topic categories and the ability to directly access databases for searches.

Still, Google Search Appliance offers a lot for large enterprises, especially on the power of its engine. This, of course, comes at an enterprise price point: The GB-1001 appliance we tested starts at $28,000, and its price scales based on the number of documents to be searched.

With these price points, probably the biggest competitors to Google Search Appliance are the companys own search services. Google provides a number of search services, including a free advertising-supported Web service that any publicly viewable Web site can deploy.

To set up Google Search Appliance we connected a system via Ethernet cable to a special administrator port on the appliance. From this point, we were able to connect to the administrator page and preset all our networking and initial administration options. Once all this was set, we could administer the appliance via a browser on any system.

From the administration interface, we could create collections, which are essentially searches of sections of a site or unique areas of content. In the most basic setup, we could define which URLs the search engine should crawl and which content and files it should ignore.

Although this was simple to do, it was not the friendliest interface weve ever seen, working mainly like a configuration file exposed in a browser form field. We could define a number of other parameters—including when to schedule crawls, the client load the crawler should use and secure password options—which now include the ability to scan Secure Sockets Layer-protected Web pages. A good number of reports and logs on usage and activity are also available within the administration interface.

Once we defined all the necessary parameters, we could run our initial crawl of the content. In tests, we were impressed with how quickly Google Search Appliance could crawl and index Web pages, completing a crawl of thousands of Web pages in several minutes.

We had some basic control of the look and feel of returned results from within the administration interface, defining options such as company logo and amount of data shown. For more advanced layout configuration, companies need to define Extensible Stylesheet Language Transformations style sheets for the result pages.

Within the search results, users get all the features theyd expect from a Google search. Besides very good ranking and document-type conversions, we also appreciated the cached content options and the support for multiple languages.

East Coast Technical Director Jim Rapoza can be reached at jim_rapoza@ziffdavis.com.