Performance Engineering Practices Improve Systems Performance by 200% in 6 Months

Performance Engineering Case Study Results 

• 200% performance improvement in critical cases in 6 months
• Fully automated testing allows for over 300 test executions a day 
• Issues are flagged at the beginning of the software development life cycle 
• Improved information analysis 
• Tests run in isolated performance environments before being deployed to production
• Time and money savings achieved by timely performance analysis of prototypes

Executive Summary

Perficient Latin America's client, a leading provider of integration software, achieved improved and consistent performance in complex, high-load products by deploying a team of highly skilled nearshore performance engineers. The nearshore performance team increased the number of test executions, enhanced analysis processes, and improved system performance across the board by applying data-driven insights from automated testing.

Tech Stack

Load / JMeter / Jenkins / Hudson / AWS / Terraform / ElasticSearch / Kibana

Background

Challenge 

Solution

To achieve more accurate predictions and in-depth insights without dedicating countless hours to running manual tests, the Perficient Latin America team began by automating all performance tests using Hudson and later migrated to Jenkins. They used Terraform to create environments in AWS, ElasticSearch as a database, and Kibana to display performance dashboards. The team went from running 6 tests a day to executing between 200 to 300 performance tests in the same time.

Thanks to fully-automated testing, results were now easier to track and changes to the code safer to deploy. This also gave the technical and business teams the metrics they needed to set achievable goals and fix performance issues before production.

The Perficient Latin America team developed a way for the client's development team to test changes to the code in safe, isolated environments. It was achieved by creating environments in AWS and performing load testing using JMeter. Moreover, the strategy only allowed one change or factor to be altered at a time. This allowed them to accurately identify effects on server runtime later, saving time and cutting costs in the long run.

Under the guidance of the transversal Perficient Latin America performance engineering team, critical use cases saw performance improvements of more than 200% in less than six months. The end product, which was improved and iterated on through performance testing and tuning, resulted in the client securing larger contracts. In the future, Perficient Latin America will be working closely to develop new features for end users and continue to ensure a faster, more stable system.

___________

Ready to build high-performing applications? Meet with our performance engineering team!