A federal government agency needed to modernize a mostly COTS legacy system which holds the data for 9+ million documents so that it can be more scalable and easier to integrate and easier to monitor and test. They also needed the building and deployment process for the system to be improved.
The agency selected QAT Global to provide the path to a successful system modernization.
QAT Global’s Approach
This system was designed over 10 years ago using technologies available at the time and has been supported in production since then with minimal changes. The key challenge facing the development team was that the existing system used RMI services, which have no load balancing and are under heavy load from both internal and external clients. External systems that use the services must be able to continue to use the existing RMI services until they have funds to be modified. In addition, existing code used ant build scripts and involved manual deployments by developers.
QAT Global designed a solution to create RESTful services which covert the existing functionality implemented by the RMI services. The RMI services will be modified internally to use the new REST services where necessary. Test harnesses for invoking and validating data from the services were also created using Angular JS. CICM was implemented using maven builds run by Jenkins and automatically deployed to the JBoss CloudForms servers with Puppet.
QAT Global’s jDaptive framework was used in the development of the RESTful services. It was adapted to use some of the required frameworks such as JPA and RestEasy. The service which loads image data was developed to load into the legacy Oracle image database as well as the new CMS Cassandra/GPFS system so that both systems ran in parallel in production until the system was proven to perform and scale properly. Once this decision was made, the system had a flag set to load and retrieve only from the new CMS, providing a seamless transition.