OBX for Google Search Appliance

The enterprise search products offered by the Enterprise Search market leaders companies are mostly large software applications that can require months to customize, train and install the complex software. They can sell for several hundred thousands of dollars to several million. There is a strong vibrant industry of consultants who help companies determine the correct enterprise search solution for their needs and then help them with the installation and integration of the software. The majority of the large enterprise search companies have global sales offices and resellers who contact companies through one-on-one sales calls.

The exception to this approach to the enterprise search market is Google. Rather than offer a complex search application that takes months to install and integrate, Google has developed the Google Search Appliance (GSA) that can be plugged in and installed within a few days. Google sells this hardware directly through the web and through a network of resellers. And rather than costing millions of dollars, this appliance starts at $30,000. The GSA does not do everything and misses some of the features of the large enterprise search applications, but their approach to a low end solution and their pitch for simplicity has paid off. It is estimated that Google has sold over 11,000 search appliances to companies last year with revenues north of $300 million. Some analysts believe that Google has overtaken Autonomy as the market leader of the enterprise search market.

This “plug-and-play” approach to search allows IT managers to provide a low cost solution to their companies, and meets the demands from management when they pressure the IT department to “give us a search like Google”. The IT manager points the GSA at content within the company they want searchable and then deploy the familiar Google query box to authorized users. The search results come back in the Google search results template, which people have come to trust and value. One study had two groups of people look at identical search results. The only difference was that one set of results were branded with the Google logo and template, where the other set of identical search results went unbranded. The Google branded search results were rated as 4 times more relevant than the unbranded results, even though they were the same results.

One of the major downsides to the Google Search Appliance is that it cannot handle massive data sets. The low end GSA will only index and search 500,000 documents at a cost of $30,000. Their largest configuration will only index and search 30 million documents. Google charges $1,500,000 for this product and it comes in a rack of 12 servers. If a company has tens of millions of documents, the price tag to do search across all of these documents can run in to the millions of dollars. One specific problem for companies is that Google counts each record in a database as a document. So a company with decades of customer transactions and sales data within a database could easily see database sizes with tens and hundreds of millions of records. Since a large database chews up so much of the document capacity of a Google Search Appliance, most companies have not included their databases in their searchable content of the GSA

Because of our technology innovation, we are especially deft at handling massive data sets. On a server that is configured very similarly to a GSA, we have been able to index and query data sets with close to a billion records on a single server. We can easily index and search 100 times more content on the same box as the GSA can.

We have four stages in our approach to the Enterprise Appliance market:

• Google Search Appliance Database Extender. By utilizing Google’s API calls and their open developer tools, we can piggyback Google’s brand and market share and offer an appliance that sits alongside the GSA. Our Database Extender allows Google Search Appliance customers to access their large databases, index them, and make them searchable through Google’s search process.
  Perfect Search sells this Database Extender Appliance primarily through the existing Google reseller network, as all of these resellers have existing customers that have the need to open their databases to search.
• Database Search Appliance. There is only a small amount of engineering and UI work to make the GSA Database Extender appliance a stand-alone application. With this Perfect Search Database Search Appliance, we can offer the major database platforms an OEM deal for a specialized search product for their database applications.
• Google Search Appliance Email Extender. Just as companies have not been able to utilize the GSA, as a cost-effective search for their very large databases, companies have also avoided using their GSA to index and search their email archives. In this day of litigation, companies are required to store all of their email correspondence. This can create massive data archives of millions and millions of emails. One of the fastest growing segments of the enterprise search market is the email e-discovery business where the major enterprise search competitors are providing solutions to index and search these large content repositories of email archives.
  In the first quarter of 2009, Perfect Search will release a Google Search Appliance Email Extender that uses the functionality and branding of the Google Search Appliance and opens up the email archives with an appliance that can handle hundreds of millions of emails on a single server. The main engineering work consists of building simple interfaces between our appliance and the major email applications of Outlook Exchange, Lotus Notes and Domino, and Groupwise. Again, this Email Extender will be sold through the existing GSA reseller network.