This is something we see often; internal systems built years ago, under intense conditions and without any long-term structure in mind. These systems are not made for upscaling and are not cost effective, which means that they need to be replaced sooner rather than later.
Our mission was to restructure the client's internal system, a system used by 200 employees every day, and doing so without disrupting their work. We developed a new architecture and worked closely with the client's own team throughout the entire project to find a way to replace the system without affecting operations, and at the same time ensure future scalability.
Replacing the system piece by piece
Because of operations and the fact that the application could have no downtime, it was not possible to build the entire system and then simply replace it. We had to construct and implement the application piece by piece. In order to do that, we created a system where we could activate and deactivate the different components, which was something the client requested during our discussions. This made it easy for us to switch between old and new system components.
If we tested a new component and it didn't quite work, we could quickly go back to the old one without it affecting the users. When we had solved the issue, we activated the new component again and made sure it was stable. This way, we could develop the new system systematically in a safe, scalable and structured way.
–We worked closely with the client, and made sure the work didn't cause any downtime, says Eric Lavesson, senior architect. He continues: However, we also paid a lot of attention to the new application's structure, as that is our main area of expertise. The client needed a scalable structure that prevents them from ending up in this situation again.
Involved Edumentors
Eric Lavesson Mattias Andersen