A large insurance group’s search for the right change data capture (CDC) software and data replication capabilities forced a showdown between Information Builders' WebFocus products and data management software from Attunity Ltd.
Montpelier, Vt.-based National Life Group is a collection of life insurance, annuity and fund management companies. About two years ago, the group decided it was time to start better integrating data from its various legacy systems and other information sources. It was clear right away that this was going to be a tough job, according to Jarugumilli Brahmaiah, enterprise data architect with National Life.
Brahmaiah said that accessing data and producing business intelligence (BI) reports had always been challenging at National Life, mainly because of the organization’s highly heterogeneous IT environment. The group’s primary data is stored across a wide array of data sources, including mainframes running IBM DB2, IBM Virtual Storage Access Method (VSAM) files, and Microsoft SQL Server databases.
Brahmaiah said getting those disparate systems to talk to one another was a difficult task, and producing critical BI reports for business users required highly labor-intensive and expensive custom programming. National Life needed to update its approach to data management. To accomplish this, the organization launched what it calls the “Data Initiative,” which is designed to help integrate and deliver accurate, consistent data to business users in as close to real time as possible.
To get the data initiative under way, National Life decided that it needed change data capture software and data replication capabilities to enable operational business intelligence (BI) and a large-scale master data management (MDM) project.
Change data capture software has been spotlighted in recent years as a reliable way for enterprises to relay information changes to databases and data warehouses in near real time. The technology identifies changes made in one data source and then makes use of data replication technology to deliver those changes to any other enterprise data repositories.
For example, a change data capture tool could take new customer information in a database connected to a retail checkout computer and then replicate that new data directly into a BI system. In a fast-paced world, the technology ultimately helps ensure that business-critical decisions are based on the most reliable and up-to-date information possible, according to industry experts.
A CDC software evaluation
National Life quickly narrowed its choice of software vendors to two candidates: New York City-based Information Builders’ WebFocus tools; and Burlington, Mass.-based Attunity Ltd.
“We looked at [Information Builders’ WebFocus tools],” Brahmaiah said. “But it was not [seamlessly] integrated with SQL Server.”
The insurance group chose Attunity primarily because it could be installed and managed as a “layer on top” of SQL Server, he explained. National Life went live with the CDC software and replication technology in June 2009, and the company says that Attunity has so far been a big help in terms of the organization’s data integration and MDM efforts.
“What Attunity does is it logs the [changes] for various activities,” said Dan Detterman, another enterprise architect with National Life. “For VSAM, you can turn on logging and, as somebody updates a policy with a status code change, [Attunity] captures that policy as being changed and then writes it to a log file.”
As a result, Detterman explained, National Life’s recently implemented data warehouse reflects the information found in source systems in near real time.
“Now we can push near-real-time data out to the reporting layers,” he said. “If [business users are] trying to track a policy and see if it’s changed its status today or in the last hour, we can now present that [information] and report on it.”
Attunity targets heterogeneous data sources
Attunity is a small and not very well-known software company, but it boasts a loyal base of high-profile customers such as Lloyd’s of London and AccuWeather.com as well as high-profile business partners like Oracle, and Microsoft, according to Merv Adrian, founder and principal of IT Market Strategy, a Pleasanton, Calif.-based consulting firm that focuses on BI and data management.
“They have great strengths in heterogeneous data sources, especially relatively obscure but widely used legacy ones,” Adrian said. “And they have solid [CDC software].”
Adrian said that Attunity’s small size could potentially be a challenge, simply because of the large number of data sources it maps to and the many products it provides. But he added that it’s likely that Attunity avoids size-related problems mostly by providing technology that doesn’t require frequent software changes or technical support.
“They aren't trying to rule the world, and I suspect they will be around for a long time to come,” Adrian said. “Their niche is one that will always be around, and the barriers to entry are high for any potential competitor [that] hasn't gained the [same] depth of understanding [of] legacy-based and heterogeneous data.”
MDM and BI lessons learned
Currently about halfway through the data initiative, Detterman and Brahmaiah say the experience has taught them valuable lessons that could benefit others.
Detterman pointed out that while it’s important to break down information silos, it’s a good idea to look at those silos holistically and come up with a standardized data model rather than dealing with each silo individually. It’s all about “harmonizing the data,” he explained.
The enterprise architect said the simplest example of harmonizing the data involves insurance policy numbers. The different systems at National Life each construct policy numbers in a different way. But rather than going around to each siloed system to change naming conventions, Detterman and his team created a cross-reference table which states that the system which originates the policy dicates the official policy number going forward. Any BI systems can now access that cross-reference table to make sure that the right policies are being used when creating reports.
Brahmaiah said the project has also thrown a spotlight on the idea that the systems containing primary “master data” and BI systems do not necessarily have to be logically dependent on one another.
“The only dependency is physical in terms of data extraction from source systems,” he said. “We extract our data off our source systems, we build something called enterprise staging, and enterprise staging is now a platform which [data sources] or BI systems can [use]. It’s like twisting the tail.”
Brahmaiah added that from an MDM perspective, the benefits of this approach have been especially clear.
“The benefits are that you have clean information in one place,” he said, “so your communications are easy and you are able to hit your target well in terms of campaign materials, marketing materials and risk assessment.”