The client had a legacy trading platform that had grown and evolved over time. The platform consisted of a stack of 33 applications, built on a variety of technologies and architectures.

Testing new features and additions was proving to be a big challenge. A simple change in one feature would warrant a verification of the complete application. To ensure that any change does not affect other functionality, the client needed to do extensive regression testing and verification.

This was a cumbersome process with over 20,000 or 30,000 test cases being checked and executed manually. The trading firm had to deploy over 20 people to carry out this exercise. The client had tried to automate the testing process with a variety of tools but was not able to get the efficiencies it wanted.

In addition, the client had multiple squads working on different apps, functionality and features. Each squad used its own automation suite. It was becoming a challenge to co-ordinate the work of the different squads and ensure that changes made by a squad did not impact the overall functionality of the platform. Iris’s brief was to design and deploy a common cloud-based test automation framework for the client’s trading platform to ensure that it could launch new features faster.

Using its cloud-based ready-to-deploy test automation framework, Iris sped up the deployment of new features for the client’s trading platform. The cloud solution, based on Amazon Web Services (AWS), featured continuous testing of multiple products on a common framework layer. It allowed for complete capacity planning of spinned cloud instances and need-based shutdowns.

Iris executed the project using acceptance test driven development (ATDD), a methodology that involves collaboration between customers, business teams and development teams. The teams jointly created the user stories and put down the acceptance criteria for any feature or functionality. Then tests were designed within the common framework to check if the feature met the acceptance criteria.

What was unique about the approach? Typically, automation is introduced towards the end of a development cycle. You would find that, in most projects, developers bring in automation in Sprint 4 for features developed in Sprint 1, 2 and 3. As a result, return on investment isn’t maximized. Our team introduced ‘in-sprint’ automation, enabling 90% test automation with every sprint. This resulted in more efficient and faster testing, and cost savings for the client.

The client’s deployment speed improved significantly with 90% faster execution in each sprint cycle and 80% faster script development.

The cloud-based solution is 100% configurable for on-demand execution on AWS, which reduced the client’s cloud infrastructure costs by 70%.

The new ability for complete capacity planning through the use of infrastructure-as-code (IaC) for spinning up cloud instances helped the client achieve end-to-end (E2E) automation of regression/ functional test cases.