Analysis of data virtualization & enterprise data standardization in business intelligence

Business Intelligence is an essential tool used by enterprises for strategic, tactical and operational decision making. Business Intelligence most often needs to correlate data from disparate data sources to derive insights. Unifying data from disparate data sources and providing a unifying view of data is generally known as data integration. Traditionally enterprises employed ETL and data warehouses for data integration. However in last few years a technology known as “Data Virtualization” has found some acceptance as an alternative data integration solution. “Data Virtualization” is a federated database termed as composite database by McLeod/Heimbigner's in 1985. Till few years back Data Virtualization weren’t considered as an alternative for ETL but was rather thought of as a technology for niche integration challenges. In this paper we hypothesize that for many BI applications “data virtualization” is a better cost effective data integration strategy. We analyze the system architecture of “Data warehouse” and “Data Virtualization” solutions. We further employ System Dynamics Model to compare few key metrics like “Time to Market” and “Cost of “Data warehouse” and “Data Virtualization” solutions. We also look at the impact of “Enterprise Data Standardization” on data integration. Thesis Advisor: Stuart Madnick Title: John Norris Maguire Professor of Information Technologies, Sloan School of Management and Professor of Engineering Systems, School of Engineering Massachusetts Institute of Technology