Entry-level: Usually starting at less than $20,000, these systems typically have limited or no security, can index HTML files and file systems with common business documents, and have vendor-imposed limits on capacity and indexing speed. They also typically do not support high availability, federation across multiple indexes or features like clustering. Examples include Google Mini, IBM OmniFind Yahoo Edition (free), Microsoft SharePoint 2007 for Search, the open-source Lucene and some Isys Search tools.
Midlevel: These systems typically start between $20,000 and $80,000 for single server configurations, and offer greater capacity, high-availability configurations, security, clustering, extensive language support, deeper connectors into enterprise applications and federation. Examples include Google Search Appliance and tools from Coveo, IBM, InQuira, Isys Search, Microsoft, Oracle, Recommind, Vivisimo and ZyLab. SAP plans to offer its own tool this year. “For basic full-text search, there’s not terrible differentiation among them—but at this tier you start to see vendors excel in particular areas, like customer self-help applications in InQuira’s case, and ease of setup and administration in Vivisimo’s case,” Brown notes.
High-end: These systems typically start at $80,000, averaging several hundred thousand dollars and extending into millions of dollars for very large-scale systems. But they provide massive scalability, structured data search and data cleansing tools, “fuzzy” search to identify related content and handle ambiguous terms within the search context, taxonomy management, data visualization tools and integration with many back-end systems such as database, ERP and CRM systems. Most of the companies offering such tools have been around for many years, including Autonomy, Convera, Endeca Technologies and Fast Search & Transfer. IBM has also recently entered this market.
Special-purpose: These systems are not distinguished by price or scale but by domain specificity. For example, Nexidia can search audio files, say, for the use of competitors’ names in customer service calls, to see if a competitor is gaining traction among customers. Fios has crafted a search engine for legal discovery.