Apr 12, 2023
Application Benchmarking for US Medical Imaging Application in Clinical Trials Space
Client is a U.S. based leading Clinical Trials organization with a focus on pharmaceutical and medical device innovation. Their primary focus of the client was on medical imaging for clinical trials.
The client had a clinical trials product in the medical imaging area, which was being used to capture, collect, control quality, provide medical imaging service to the final delivery to the sponsor company and to the FDA or EMEA if required for medical images in a wide range of imaging modalities.
Client had done a technology upgrade to the product and wanted to benchmark their new platform with the existing system
New platform dealt with medical images needed to be assessed had made use of Java Network Launch Protocol.
How did we help the Client?
We worked closely with the platform SMEs, business analysts to identify a list of end of end business critical transactions, identify business usage patterns to develop a workload model to simulate real-life load on the system.
The old platform’s thin client performance was benchmarked for the thin client making use of Jmeter tool. Later the thin client performance for new platform was conducted with varying concurrent user load and the most optimal load the system could handle in current design was identified.
The new platform’s thick client had complexities involved using JNLP, the NeoLoad tool was brought in place of Jmeter for thick client, as the plugin support for thick client wasn’t good enough.
Load, Endurance and Stress tests were carried out to measure the performance of the server and the bottlenecks in various layers.
Load tests unearthed 8 key performance issues across database layer and application server layers, brought out distinct insights to performance improvements in areas of search operations, imaging operations. Performance improvements helped meet < 8 seconds SLA, which in turn provided an assurance for customer satisfaction.
Certain recommendations were provided on the server capacity sizing, to leverage optimal resource usages, which was acknowledged by architect and infrastructure teams. Capacity recommendations resulted in higher volume of operations being handled with 300 concurrent users, than earlier.