Our customer had developed a custom web application to create and manage quotes from different vendors, view and search properties within a single platform for wastewater management services.
The system was supposed to scale up to a total of 500 users during peak hours. The expected response time was 5 seconds for both transactional and non-transactional pages.
The system was expected to scale up dynamically at runtime if the load is increased. But, later on, it was noticed that there was a remarkable degradation in application performance resulting in low page responsiveness and page load time with increased load. It had a noticeable impact on business goals giving rise to idle resource time, less throughput and large turn-around times.

In order to overcome the bottlenecks, ensuring seamless performance, the customer needed a best-suited solution within the stipulated timeline and budget. Keeping in view the time and budget constraint both the parties agreed upon with the implementation of open source tools.

Our first step was to decide on the right tool, amongst the wide range of tools available in the market that could boost application performance and client confidence. In order to shortlist the right tool, the above aspects were taken into consideration.

On the basis of the above, based on feature analysis, we shortlisted open source tools like Pylot, OpenSTA, and JMeter.

jMeter was chosen as the best choice based on our experience with other tools and current context

• Support and community for jMeter are very active when compared to any other open source tool. jMeter is Java-based and easy to learn. Java being platform independent, the tool is cross-platform execution ready
• jMeter is based on the simple record, customize and playback concepts. Thus, testers are not required to write functions for basic features such as content check and correlation
• As GUI is very user-friendly, recording application sessions and executing them is very easy. Third-party Amazon image for cloud-based testing is also available
• jMeter has the ability to distribute the generation of load across multiple load-injector machines (i.e. jMeter supports multiple load-injectors managed by one controller)
• HTTP header cookies are managed automatically in jMeter and can be manipulated manually if required

By executing a well-planned QA process and strategy, we helped the client attain the following benefits:

• Response time improved to 80% as compared to the existing application
• Reduced risk and overall time for testing life cycle
• Match the performance benchmark that led to the client’s (team’s) confidence in the software.