|
|
Key Technology Differentiators
BA-Insight for Microsoft SharePoint,
Search Server and FAST
Home > Technology > Key Technology Differentiators
| Key Technology Differentiators |
Leverage existing infrastructure: BA-Insight leverages infrastructure in two ways. First of all, it connects heterogeneous backend systems into unified information architecture. Second, it leverages the existing SharePoint or Search infrastructure, by deploying a thin software layer on top. No need to additional support personnel, disaster recovery or scalability planning, which are significant drivers to lowering the Total Cost of Ownership (TCO).
|
Real-time vs. Cached Information Integration: The caching of structured and unstructured data is pivotal to Unified Information Management and delivering a single point of access to all enterprise data. Because data, content, security, and usage are stored in disparate proprietary formats and systems distributed across the enterprise, it is imperative to store all these attributes in a single location and in an open format for the purpose of unification. This principle is well-known of 30-year data warehousing best practices, yet many information integration efforts will attempt at by-passing such principles by implementing direct API calls into the disparate source systems. The result is an architecture which will fail at scaling as the workflow bottleneck will systematically be the weakest transformation or connection in the pipeline.
|
For example, let’s assume we are indexing content in Lotus Notes and need to map security information from Lotus Notes to Active Directory. The actual lookup of security information while indexing content from Lotus Notes creates a significant overhead to the speed of indexing, resulting in a significant duplication of lookups of the same user credentials across all records. By contract, a cached Information Integration architecture will map security information as a separate job once. When the content from Lotus Notes databases is being indexed.
|
Real-time vs. Cached Information Processing: Enabling asynchronous parallel processing also enables rapid scaling, without creating dependencies between the various information processing stages. Take for example the calculation of a document preview and concept extraction in real-time. Such action will occur over and over for a highly requested document instead of being processed and cached once. Given the bottleneck and costs associated with CPU resources vs. storage resources, it does not appear to be practical. Moreover, the time that it takes to compute the preview and concept extraction negatively impacts the user experience, as the alternative is simply to open the native document and benefit from the native application context to perform the document analysis. Worse, if a document is local to the user, then the network latency related to sending the document to the remote processing servers needs to be added to the wait time. In an Enterprise Search or Business Intelligence context, fast query access is primordial, which is the reason for cached index or data warehouse architectures.
|
Instant and Actionable Information Access: Users want instant access to information, especially in the context of Enterprise Search and Business Intelligence. Pre-processing and caching of information insights is the basis to deliver instant information access in a consistent and reliable fashion. Users also want actionable information access. A rich document viewer built in Flash or Silverlight Rich Internet Application (RIA) frameworks is the only viable option to deliver actionable insights and context to the user. The use of raster images for preview for example will result in large and static file sizes, which cannot be resized without re-processing the information or have text selection. By contrast, an RIA framework such as Silverlight enables resizing, selection of text, and many more actions, including custom .NET actions that can be built right into the Silverlight package.
|
Automatic Relevancy Feedback: AptivRank is a unique patent-pending technology that rivals Google PageRank™ but for Enterprise. It is a reliable mechanism used to track search result item popularity based on explicit user feedback built-into the search workflow. An explicit information consumption action, such as print, download, or share is recorded as a vote of usefulness of the information at hand. The more votes a piece of information gets, the greater the AptivRank score, which is fed back to the relevancy ranking algorithm.
|
Unified Information Store: The Enterprise Dataspace™ platform is built using XML and web service standards to enable out-of-the-box interoperability with existing infrastructure and provide end-users direct access to structured and unstructured data, regardless of proprietary formats or repositories.
|
|
|
|