Next Generation Business Intelligence
Thursday April 12, 2012
11:00 a.m., 1000 SEO Building
Business Intelligence (BI) refers to technologies for collecting, integrating, analyzing, and presenting large volumes of information to enable better decision making. Traditionally, BI systems have been designed to support off-line, strategic decision making. As enterprises become more automated, real-time, and data-driven, BI systems are evolving to support on-line, operational decision-making. In this talk, we describe the challenges in designing and optimizing this next generation of ?Live BI? systems. Many of the challenges stem from the need to scale to enormous data volumes (the ?big data? challenge), the need to integrate a variety of new data types (semi-structured and unstructured data, live streaming data from sensors and social media channels) into the pipeline, the need to enable near real-time decision making, and the need to trade off various quality objectives such as cost, latency, and fault tolerance. In the modern BI architecture, the back-end data integration flows and the front-end query, reporting, analytics, and visualization operations are fused into a single analytics pipeline, where analysis may occur at any stage in the flow. Instead of a “one size fits all” engine, there may be many choices of engine to execute different parts of the pipeline: column and row store DBMSs, map-reduce engines, stream processing engines, and analytics engines. We discuss promising approaches we are taking to meet these challenges.
Umeshwar Dayal is an HP Fellow and Director of the Information Analytics Lab at Hewlett-Packard Labs, Palo Alto, California. Umesh has over 30 years of research experience in data and information management. His current research interests are in enterprise-scale information management, live business intelligence, data mining, analytics, and information visualization. Prior to joining HP Labs, he was a senior researcher at DEC’s Cambridge Research Lab, Chief Scientist at Xerox Advanced Information Technology and Computer Corporation of America, and on the faculty at the University of Texas-Austin. He received his PhD from Harvard University. He has published over 200 papers and holds over 50 patents. Umesh is an ACM Fellow, and the recipient of the 2010 Edgar F Codd Award from ACM SIGMOD for his contributions to data management. In 2001, Umesh and two co-authors received the VLDB 10-year Best Paper Award for their 1991 paper on a transactional model of long-running activities. He is on the Editorial Board of several international journals, has edited two books, and has chaired and served on the Program Committees of numerous conferences. He is currently a member of the Steering Committee of the IEEE International Conference on Data Engineering and the ER Conference on Conceptual Modeling, and has served as member of the Board of the VLDB Endowment, the Board of the International Foundation for Cooperative Information Systems, and the Steering Committee of the SIAM Data Mining Conference. Umesh can be reached at firstname.lastname@example.org .