In-Depth

Acquisitive ING links silos

• By Lana Gates
• March 28, 2003

ADT's 2003 Innovator Awards

Data Warehousing Finalist

ING U.S. Financials, Hartford, Conn., offers services in life insurance, annuities, retirement plans, employee benefits and mutual funds. The company has expanded rapidly through acquisitions, resulting in vast amounts of data being locked in disconnected silos, many of them located on proprietary legacy systems. In an effort to integrate data from more than 100 heterogeneous and legacy systems, ING developed the Enterprise Information Infrastructure.

Part of the project's goal was to move from a batch-processing model to a near real-time model to improve business results and comply with federal regulations. The project focused on operational data delivery and analyzing the data for optimal business performance.

With more than 5,000 users and millions of dollars worth of transactions processed by the system daily, the project posed many challenges. ING's largest risk was that errors or delays in the project would thrust the company out of compliance with SEC and NSCC regulations, a failure that could cost the company millions of dollars.

The development team had to reconcile the long-term goal of creating a strong, flexible and scalable infrastructure with the short-term need for return on investment and servicing immediate business needs. The team employed a top-down, bottom-up approach to design and implementation.

From a technological standpoint, the project required the design of a uniform information infrastructure flexible enough to incorporate 14 separate financial service operations with different technical architectures into a single system capable of processing data in real-time. The development team met this challenge through intensive research, dedicated experimentation and appropriate use of resources.

In addition, some of the team's original assumptions regarding existing systems were incorrect - for example, data fields they thought existed did not or they did not match up with the fields from other systems.

The outcomes of the project, which was implemented in a period of 10 months, include consolidated statements, cost reduction in compliance-required retail mailings, consolidated resources, regulatory compliance, fraud detection, and improved integration and processing of new assets.

The successful implementation of the project also resulted in the best practice of creating an enterprise model based on business function process. ING gathered all of the business people involved into a conference room for three days and had them help to define business function, according to Babu Kuttala, director of information management. The company then created its enterprise model based on those definitions, according to Kuttala.

The project's success has paved the way for expansion into new source systems, interoperable meta data management as well as additional personalization.

If ING had the project to do over again, it would allocate more time to data quality issues and put a corporate-wide data quality initiative in place before starting. The company would also add more source system expertise earlier in the process, according to officials.

Application profile

Project: Informatica data integration project

Purpose: To integrate more than 3.0 terabytes of data from 100+ heterogeneous and legacy systems.

Benefits: Unified customer view, improved customer service, more effective marketing campaigns, improved sales reporting and analysis, improved business management.

Platform: Sun Solaris

Tools: Informatica 5.1 PowerCenter, IBM MQSeries, JMS, legacy VSAM databases, Qualix, iPlanet Web servers, WebLogic

Development team: Senior manager, project manager, technical architect, database design lead, operational DBA, ETL lead, four ETL developers, three business intelligence/analytics specialists, two business analysts, two QA resources and a documentation specialist