
Wrapping the API' s (POC)

The Problem
Queensland Department of Transport and Main Roads (TMR) was in need of modernising a large legacy system. The legacy system was on a dated mainframe with significant costs. The strategy was to implement an anticorruption layer (facade) around the legacy system and then gradually replace it with micro-services. There are 100's of endpoints that require implementation. A pilot showed that it would take 3 to 5 days to implement each endpoint. With hundreds of these to implement, the time would simply blow out and costs would be too high.
The Solution
TMR commissioned a pilot to investigate using automation on the project and approached WorkingMouse.
The WorkingMouse approach is to apply the principles of Jidoka (automation with a human touch) by using bots to code the majority of what would have manually been done.


Automation at Scale
This is achieved by using a Codebot. A Codebot enables developers to fashion specific software engineering tools to achieve high-levels of automation, in other words, automate software development.
The Outcome
TMR's new custom built bot used the WSDL descriptions of the APIs to generate over 90% of the code that was being written manually.The time taken to implement an endpoint was reduced from 3 to 5 days down to 1 to 2 days. Conservatively, it is an estimated 60% saving.
With the high-levels of automation, TMR is now able to implement 100' s of endpoints and continue their large-scale modernisation.
