On-demand KM: a two-tier architecture

On-call forms of KM generally require that certain KM elements, depending on the application, perform their functions in anticipation of need. How far in advance this preparation occurs depends on the application. In general, when the preprocessed information's range of application is broad, there is little advanced notice of need, and the need is immediate and on-demand, a form of batch processing is appropriate. Other KM elements, such as those for synchronous, real-time collaboration, require no preprocessing. Also, the KM portal that hosts the KM elements seldom needs preprocessing. Thus, for on-demand KM applications with anticipatory processing and a real-time front end, two or more tiers working independently are appropriate. Performing some KM functions in a continuous batch mode and others in real-time can make on-demand applications practical.