SaaS-based ERP Application Enhancements and Upgrades

The ERP application was already live for some time. But, the client started seeing loopholes in the app’s technical architecture. The loopholes were primarily because of poorly-written rule engines.

Rule engines are libraries that pull the business logic out of a system’s core structure. That is, they decouple logic from an application’s core. This decoupling helps developers make rapid changes in business rules/logic without affecting the application as a whole.

But, our client’s existing ERP app had badly-written rule engines. Many of these rules were not appropriately responding to specific events, user inputs, or actions. So, our client wanted a highly-skilled development partner who could do the fixes and define the rules correctly.

After going through Capital Numbers’ portfolio, the client found a trusted partner in us. So, he reached out to us for help.

Our primary challenge was to understand the flaws in the existing rule engine designs. We then had to come up with deep-nested logic. We had to skillfully match algorithms to define which rule should run when and in what order. The work required us to develop several intelligent ‘if-this-do-that’ conditional statements.

Our next challenge was to fix the caching. We saw that the existing cached memory didn’t clear off data when users logged out. As a result, the app would load slowly. So, we had to offer solutions that improve the data retrieval process and clear off cache data when users log out.

Our third challenge was to upgrade the frontend framework. We had to add new UI components like menus, lists, boxes, drag-and-drops, etc. A version upgrade was, therefore, crucial.

After noting the key challenges, we began the project with an as-is analysis. We checked the deep-seated issues and implemented the following solutions:

Firstly, our examinations confirmed multiple areas were error-prone. So, we removed those errors and wrote powerful rule engines in the existing .NET Core backend environment. We resolved complex, underlying issues using Data Structure and Algorithm (DSA) techniques. We also double-checked algorithmic efficiencies after implementation.

Next, we created a cache memory store. We did this to help the app fetch user-requested data from the cache (a quicker process) and not from the main memory (a lengthy process). We also ensured the cache store clears temporary files when users log out.

We then worked on the app frontend. We upgraded it from Angular v10 to Angular v14 because the latter offers a variety of reusable templates and custom page title options. The v14 also helps customize menus, tables, overlays, dialogs, drag-and-drops, etc. So, upgrading to v14 was essential.

In addition, we used Syncfusion to design UI elements on a granular level. Syncfusion is an excellent tool for creating data tables, grids, charts, graphs, etc. We also build logic to connect the frontend with the backend data.

Next, we used Docker Compose to containerize different microservices (Finance, HRs, Customers, etc.) within this single ERP solution. Containerization puts less load on servers. We also used Kubernetes to manage the individual containers (microservices) efficiently.

Our experts used two database solutions for the ERP app. We chose MSSQL because it allows excellent data restoration and recovery. We also used MongoDB because it scales well and has outstanding data-querying ability.

Once we implemented it all, we did unit testing using xUnit.net to check if the backend functions were running correctly. After a thorough quality check, we deployed the app in MS Azure, as Azure is excellent at processing enormous data volumes.

The outcome is an improved ERP app that is more workable and functional. The app is better equipped to handle CRMs, Sales, Accounts, Finance, and HR tasks. Here’s more on the final results: