Optimal enterprise data architecture using publish and subscribe

In medium-big enterprise it is quite typical that the database architecture is defined through a sequence of projects and realizations that result a number of different and sometime overlapping data sources. This trend is worsened by merger and acquisition activities that add in existing data archi-tecture new data sources from external organizations. Data fragmentation reduces significantly the possibility to exploit organizational information assets and it needs the building of a data integration architecture. This architecture allows the organization to access to data stored by heterogeneous data sources and to manage updates through a unified view of this data. In this paper we present an original framework able to support the evolution of data architecture by identi-fying the optimal solution that maximize the quality of the overall architecture within a given cost threshold. In partic-ular, we focus on Publish and Subscribe as solution for data architecture. Copyright 2010 ACM.