Insurer soothes its data accumulation headaches

One insurance firm says it expects active archiving software to save it money and improve database performance, but that privacy compliance will likely require additional tools.

For several years, AXA Assurances Inc. in Toronto was using a “fairly simplistic” toolset to purge and archive data whenever necessary, said Debbie Smith, vice-president of development at the company’s technical division. That tool came with software from Policy Management Systems Corp. (PMSC) the firm was using to manage the information stored in its production databases.

But in 1997 AXA switched to a new DB2 application system, Huon, which did not come with similar tools. At the time, excessive data on the production database was not an issue. But over the years it built up to the point where AXA now has “close to 300,000 policies on the database. That’s millions of rows of data,” Smith said. With the database growing at a substantial rate, “it was becoming critical that we address the issue before it became such a problem that we were behind the eight ball.”

There were two ways excessive data could cause problems. The first one was mainframe costs. “We have our mainframe processing done at our sister company AXA Financial in New Jersey,” Smith explained. “They charge by the CPU minute, so the more CPU you use, the more you pay for it.”

The second reason was to improve speed and efficiency. “As the database grew, we knew it could have some degredation in performance,” she said. After estimating in-house development costs, AXA did some research and happened upon Princeton, N.J.-based Princeton Softech’s active archiving product.

Jim Lee, Princeton’s vice-president of product marketing, said the tool allows customers to identify data that is not used on a regular basis and move to a secondary, lower-cost storage device such as an online archive database, tapes or optical devices.

The user creates an access definition – the business rule that dictates what data should be archived. For example, said Lee, an insurance company can specify that claims more than two years old or information on closed cases or former customers should be archived. For the data to retain referential integrity, the definition should include “not only claim information but also data that makes the claims whole and complete,” said Lee – things like billing and contact information tied to customers’ policy records.

The software interprets the rules and converts them to SQL queries which access the appropriate data and move it to the secondary storage device on a regular basis.

Smith said AXA also considered IBM’s archiving software, but in the end, Princeton was the final choice. The vendor set up a successful live-test environment. Moreover, “it was a competitve price and met our needs; it was quick to install and very easy to set up,” she said.

The implementation was planned in four components, of which AXA has recently completed the first, already eliminating 80 million rows of data. “At this stage of game, the production environment worked well with first set of purging that we did,” Smith said.

She added that AXA probably won’t be able to assess the full benefits of the system until May of this year. But preliminary estimates suggested that the payback would be in about two years, Smith said.

Because the project is happening in stages, AXA hasn’t yet implemented the component that would allow it to retrieve the data – for now, it can only archive information. The law requires insurance companies to keep records for seven years, but Smith said AXA’s in not much of a hurry to implement the retrieval component of the system. “We’ve found that the amount of times the business needs to retrieve something that old are very few and far between,” she said, adding that she’s got special retrieval methods up her sleeve in case something comes up.

Although the fourth stage of Canada’s Personal Information Protection and Electronics Documents Act (PIPEDA), which covers the private sector, came into effect on Jan. 1, Smith said AXA did not purchase the active archiving solution with privacy in mind. “AXA is very respectful of privacy details…and has already taken steps to ensure, even before the law became as strict as it has, that it is taking appropriate measures to secure personal information.”

Active archiving in itself will not help customers deal with privacy issues anyway, Smith said. Instead, it might be more conducive to look at Princeton’s relational suite of products that comes with a masking tool. Princeton’s Lee said such a tool would allow companies required by law to retain certain data to keep it on their systems, but would prevent access to portions of the data that should not be viewed or used under PIPEDA.

Related Download
Virtualization: For Victory Over IT Complexity Sponsor: HPE
Virtualization: For Victory Over IT Complexity
Download this white paper to learn how to effectively deploy virtualization and create your own high-performance infrastructures
Register Now