Datasource Consulting helps a leading mutual fund company improve their data integration architecture.
To boost data load performance and reestablish a supported Informatica environment, this mutual fund management company set out to simultaneously update their data integration architecture and upgrade Informatica from version 8.6 to 9.1.
The challenge with this upgrade was that this environment was complex. They run about 1,000 jobs to move terabytes of data into at least 6 different target systems across multiple projects and subject areas. Dozens of developers had to be involved in the testing and migration, while simultaneously maintaining ongoing operations.
Datasource Consulting immediately identified a Senior Informatica Consultant to serve as the Administrator and consulting advisor for the project. A series of project steps were executed in sequence to achieve the project goals:
Step 1: Address the performance challenges of supporting two data centers in different cities. To address this challenge, redundant Informatica servers were established for each of the Dev, QA, and Production environments. In other words, they created duplicate instances of Informatica in both data centers, instead of just one. This eliminated the need to move large amounts of data across the WAN from one data center to the other.
Step 2: Duplicate the Dev and QA data warehouse tables. Subsets of data were used to eliminate the need to significantly expand the required disk space. Typically, historical data was excluded in the duplicate data set.
Step 3: Set up three new Informatica servers in each data center (a total of 6 servers) with the latest version of Informatica. These servers pointed at the same source systems that the old versions pointed at, but their targets were configured to point at the duplicate Dev and QA data warehouse tables set up in Step 2.
Step 4: Coordinate nightly project migrations, first in Dev, and then in QA. These project migrations were coordinated with each corresponding team lead and their developers. A set of mappings were migrated and then manually run both in the new and existing server. Care was taken to ensure that the same source data was used by both servers which required precise data load timing.
Step 5: Use Informatica DVO to test the data load results between the new and existing target databases. This step ensured that the upgraded Informatica server loaded the data in the exactly the same way as the older version.
Step 6: Once the Dev and QA environments checked out for a given project, a production cut-over was scheduled. First, the mappings associated with a project were upgraded into the new repository. The first data load on the new server was typically accomplished manually. This allowed the team to watch for failures and anomalies. Once everything checked out, CA Autosys was modified to reference the newly upgraded server for the project, and scheduling was set.
With Datasource Consulting’s help, this mutual fund company updated the architecture and successfully upgraded Informatica for 6 projects over the period of 3 months. The company is now able to load data four times faster with the new server architecture and is on the most current version of Informatica. The project was completed with no negative business impact – a key measure of success for any upgrade project.