Comprehensive systems documentation and application discovery solution – Spring 2015 to Fall 2016
A leading US life insurance company needed to improve the accuracy and completeness of their systems documentation both to help staff with their ongoing application maintenance and development and to streamline the onboarding of suppliers for outsourced projects. The company’s IT leadership team realized that they needed to ensure their core business applications remained sustainable and cost effective, and identified four areas where an application understanding tool could benefit them:
- Make application documentation widely available
- Reduce time, effort and cost to document production applications
- Improve productivity by identifying dependencies for change impact analysis
- Map application and data relationships for a complete portfolio model
The scope of this project included the entire mainframe code base for the company’s complete insurance line of business. This entailed 90+ million lines of code across a variety of languages and technologies (COBOL, Assembler, PL/1, CICS, DB2, IDMS, Culprit, ESP, etc.) comprising some 225+ applications. A major problem was presented early on: the client indicated that they could not determine which source code belonged to which application and they wanted ib-ARM to make that determination. This was the ultimate “application discovery” challenge and a fundamental departure from typical installs where the determination of what source code belongs to what application is known in advance and becomes input to the configuration of the tool.
To prove that the tool could meet this ultimate challenge we were engaged in an extensive Proof of Concept (POC) which entailed about 20% of the entire scope. We needed to develop an iterative process to determine what source code belonged to what application. It was determined that the Job Scheduler contained the Application Id for each Job. From the Job, using ib-ARM mapping technology, the Application information could be passed on to the Procs, Load Modules, Programs, etc. used by that Job. We parsed all the source code under one application and then separated the code objects into separate applications after parsing and loading. We needed a similar iterative process for CICS which we also were able to develop.
Another challenge was to create a seamless two-way connection to Alfabet, the organization’s top-down Application Portfolio Management tool. We developed this interface during the POC; enabling the automated seeding of Alfabet with key application metrics and information flows.
The POC was successfully completed by the end of summer in 2015 and the green light was given in early 2016 to proceed with the ultimate production install for the entire portfolio. With a few custom extensions the install was delivered to production in September of 2016.
Because the ib-ARM data repository resides in the cloud, it will be easy to extend access to new areas of the business as well as to key external IT service providers in later phases of the rollout.
ib-ARM is now an integral part of this company’s development and maintenance practices. The Repository is refreshed on a bi-weekly basis.