The proliferation of data in the enterprise is leading many organizations to pry their pocketbooks wide open in hopes of finding a reliable way to search and retrieve critical information lurking in various content repositories.
Nearly 100 vendors tout solutions for enterprise search, and they often take very different technical approaches to solving the problem, according to Harley Manning, research director at Forrester Research Inc., in Cambridge, Mass.
And enterprises are waking up to the critical nature of information retrieval as search becomes a core value of business process, according to Manning. "It seems so obvious, but search is fundamental. If you can't find [it], it isn't there," he said.
For example, if customers can't find products on a Web site, the company won't sell as much. Moreover, many enterprises are realizing that effective search can lead to improved employee productivity, higher sales, lower customer service costs, Manning said.
Joining the ranks of enterprise-focused search vendors, Google Inc. this year expanded beyond its Internet domain to tackle enterprise search with an integrated hardware and software search appliance. Inktomi, AltaVista, Verity, Fast Search & Transfer, Autonomy, and others target corporate content with software designed to hook into various content repositories.
Natural language processing technology is also gaining a foothold. Typically applied to customer self-search, the technology is spreading deeper into the enterprise as players such as iPhrase, AskJeeves, and newcomer InQuira expand their offerings into more types of corporate data repositories and formats.
Google this week bolstered its Linux-based Search Appliance with increased capacity, support for more enterprise content types, and simplified Web-based administration. In addition, the Mountain View, Calif.-based company announced new customers, including Boeing, Cisco, and The World Bank.
Other enhancements include point-and-click Web-based administration, improved integration with J2EE (Java 2 Enterprise Edition) applications, simplified navigation, a customizable search results interface, support for Lotus Domino repositories, URL tracking and analysis, and spell checking.
Google's appliance approach to corporate search is blazing new trails in the enterprise software space, said Matthew Berk, an analyst for site technologies and operations at Jupiter Media Metrix in New York. Typically, appliances have provided utilities farther down the stack, for load balancing and SSL (Secure Sockets Layer) acceleration, for example.
"Software run as a device is relatively new," Berk said. "It is appealing because you plug in and away you go. You configure it with a browser rather than with an engineer."
But one possible drawback is that IT managers typically don't react well to the idea of yet another box that they can't peer into or integrate with other systems, Berk added.
"Some people view the drawback of such an approach is that you can't get under the hood to do tight integration and specific configuration," Berk said. "From what I've seen of this device, that is simply not the case."
The Google Search Appliance's capability of rendering in XML and XSLT (Extensible Stylesheet Language Transformations) offers enterprises a lot of flexibility, Berk said.
"You can do great amount of customization without having to learn the internals of the box," Berk said.
In addition, "the black box is a great advantage when you are tired of installing, integrating, maintaining, and administering software that is pretty high end," Berk said.
Similar to what Cisco did with routers, Google is looking to hit the sweet spot with a best-of-breed point solution, according to Eric Schmidt, CEO of Google.
"Everyone agrees with the appliance approach and the price point," Schmidt said. "People are not saying, 'I really want complicated software that I can fiddle with and spend large amounts of money for consultants on.' "Search and knowledge management software vendor Verity week after next is expected to roll out new search and classification software for the enterprise. The product is slated to include secure content search and retrieval, taxonomy creation, and the capability to build gateways to enterprise applications, such as ERP supply chain, and CRM.
Verity stands by the need for an extensible software approach to enterprise search.
"If it were possible to [put] search in a box and serve all the knowledge management needs of the enterprise, we would have done it years ago," said Prabhakar Raghavan, CTO of Verity. "Enterprises are complex, and need flexibility. That flexibility [that] the enterprise needs typically you won't get from a box."
InQuira hit the scene this month with a set of applications that combine natural language processing technology used in customer self-service applications with more traditional information retrieval processes typical of enterprise search systems. The applications are based on the company's Natural Interaction Engine that can understand the form of a question and how to get the desired results out of Web sites, documents, and databases, according to Michael Murphy, CEO of InQuira, based in San Bruno, Calif.
"The technology combines information from the question input, contextual information, and information about who is asking," Murphy said.
The underlying technology is based on the Cooper-Yuret Language Model, which taps a deep knowledge of language, relationships between phrases, ontology, and grammatical patterns, company officials said.
In addition, InQuira's search engine exploits a semantic index that analyzes content and data sources in terms of concepts rather than key words, resulting in richer comparisons down to the sentence level, Murphy said.
Going forward, InQuira hopes to further blur the lines between self-search and enterprise information gathering by targeting more application content.
In addition, AskJeeves this week rolled out a new self-service application that combines natural language processing, customer intelligence tools, and hooks to enterprise applications.