IBM WebSphere Information Integrator, OmniFind Edition 8.21 gets good marks overall, especially for scalability, which is excellent. The genuine enterprise search engine federates data from intranets, e-mail repositories, and databases. Scalability and configurability are compelling, but some widely used file types are not supported, and security features need to be stronger. Testing of IBM WebSphere Information Integrator, OmniFind Edition 8.21 in a San Jose, California office of IBM demonstrated that IBM uses the search engine to locate items in its databases, e-mail archives, and 10,000 IBM Web sites. Two configurations are available to purchasers: a single server system and a four-way system made up of a crawler, a parser-indexer, and two redundant run times offering client interface servers. Most clients are browsers, but interaction can also be via a Java application programming interface (API), and the latter permits a department to query search results. Results are returned to the application as Java objects. The Java API is a good choice for custom software, and includes a knowledge base search facility built into a help desk application. The crawler supports database, extracts searchable data, and also can search unstructured data. Weighting of data entails the frequency with which a keyword appears in a page, if it appears in a title or subheads, and how often it is in anchor text.