Network Computingwww.networkcomputing.com

SirsiDynix: Why Search Is Being Integrated In Applications

Posted by Paul Korzeniowski
November 06, 2009

Tags: Enterprise search

Channel: Enterprise Search, Data Center

Enterprise search is being integrated into a growing number of applications. Increasingly, vendors are selling their products to companies like SirsiDynix, which are customizing it for special vertical markets. Consequently, a product's customization features and its pricing model are becoming as important as the underlying search functionality.
    
SirsiDynix, which has 300 employees, supplies more than 4,000 libraries worldwide with automation software as well as search and discovery products. The firm has a handful of data centers around the world, located in Georgia and Utah in the US, and in Australia, Canada, and Scotland. The IT software and service supplier has 32T bytes of data storage supporting its business, and that number has been growing exponentially in the past few years.
   
A few years ago, SirsiDynix started to look for an alternative for its library search system, which has been in use for more than a decade and showing its age. "We needed a search system with a better user interface," said Jeremy Klute, product manager of SirsiDynix Enterprise. "The libraries wanted something that looked more like Google and less like a traditional card catalogue."

After examining a handful of cloud, commercial and open source enterprise search products, they quickly discovered that pricing for enterprise search systems varies dramatically. "Some of the products that we examined were hardware intensive and would have required that we deploy a large number of servers," Klute explained. 

SirsiDynix eventually chose Brainware's Globalbrain. It provided the flexibility in returning search results that they needed. In a growing number of instances, autonomous libraries work together in various consortia. Users may not have clearance for all materials, so the search engine needed to be able to restrict information access to specific user groups, which Globalbrain does.
   
SirsiDynix also needed a search system that would help librarians and library users clarify their searches. Globalbrain's Facet Stacking function narrows search results by a particular facet of information, such as author, content type or holding library. Dynamic Search Suggestions are also displayed, listing other successful searches as a user is filtering information.

SirsiDynix was not enamored with traditional key word search because users often misspell search terms, so Globalbrain's enterprise search system's fuzzy matching capabilities was an attractive feature. Here, the search system will take a term, such as "enterprise," and break it down into "trigrams" (defined as groupings of three letters: ENT-NTE-TER-ERP-RPR-PRI-RIS-ISE), which are then indexed. Consequently, the misspelled 'Tchaikowski' will return correct hits for 'Tchaikovsky'.

SirsiDynix also needed a search engine that was toolset as well as indexing mechanism. "Each library has a slightly different demographic and charter, so we (as the software supplier) needed to be in a position to address those needs," explained Klute. The Brainware product provided each library with control over the search context, look-and-feel of the system, control limits and the variety of data sources available to the user.

Related Reading

More data-center Insights