MOSAIC: A cohesive method for orchestrating discrete analytics in a distributed model

Abstract

Achieving an HLT analytic architecture that supports easy integration of new and legacy analytics is challenging given the independence of analytic development, the diversity of data modeling, and the need to avoid rework. Our solution is to separate input, artifacts, and results from execution by delineating different subcomponents including an inbound gateway, an executive, an analytic layer, an adapter layer, and a data bus. Using this design philosophy, MOSAIC is an architecture of replaceable subcomponents built to support workflows of loosely-coupled analytics bridged by a common data model. © 2013 Springer-Verlag Berlin Heidelberg.