March 6, 2007

3:15 PM - 4:15 PM

Level: Advanced

Business users are demanding more data, better analytics, and faster delivery times from their I/T organizations. And the tools to support these needs—BI, ETL, warehousing, mining, cheap storage—are readily available and no longer key differentiating factors. In fact, they may be driving behaviors that are inhibiting IT’s ability to deliver what the business needs to compete. For example, when an application is built for a particular business function, it will typically manage its data within that function’s limited view of the enterprise data model. Over time, different applications will fragment the data model, as the needs of each business area evolve and time-to-market pressures make it easier to create loosely-coupled data stores for each application. Subsequently, one or more data warehouses will be built to collect and homogenize the data silos, leading to further inconsistency between the warehouse meta model and the source data models, and requiring heavy ETL transactions, redundant storage, and more data cleansing…not to mention higher development and maintenance costs. But if “the business is the business”, i.e., the business model is the same regardless of how its systems are structured, then why should each application need its own data model, and why should the data warehouse need a different meta model than the operational data stores? ESB (and more broadly, SOA) provides a framework and tools to help enforce better alignment of the enterprise data architecture across business applications and warehouses. In this presentation, we will discuss how ESB can be used to bring together multiple application data stores and data warehousing platforms under a common data access umbrella.

Three specific implementations will be discussed:

1. Simplified data access services to replace localized (app-specific) services and eliminate cumulative calls and transformations;
   
2. A common inquiry gateway for applications to navigate both operational data stores and a data warehouse;
   
3. Right-sized access services that better fit specific data requirements and reduce the need for an application to “crawl the data model”.

David Koenig is the Former Chief Information Officer, Personal Market, for Liberty Mutual Group. Prior to Liberty Mutual, he was a Managing Director at Citigroup, where he spent time as CIO for Travelers Life and Annuity, and also as Sector CTO/Chief Architect for Citigroup Global Investment Management. His work experience includes four years each at The Boston Consulting Group and National Instruments. He has an MBA from Harvard Business School, an MS from the California Institute of Technology, and a BS from the University of Connecticut.