The US’ #1 Retail Bookseller Reduces Response Times by 46%

The Client

GeneXus is a software company that conducts in-house research to develop state-of-the-art proprietary technology in the fields of database design and maintenance, application development and maintenance, and, in particular, automatic knowledge management.

Founded in 1988, the company is headquartered in Montevideo, Uruguay, has offices in the United States, Mexico and Brazil, and is present in more than 35 countries through its network of distributors.

GeneXus™ is its main product, an intelligent tool for automatically creating, developing and maintaining mission-critical, multi-platform applications.

In 2017, Genexus™ was named a new challenger by leading research and advisory firm, Gartner, in its App Development Platforms Magic Quadrants Report.

The Need

The largest retail bookseller in the US commissioned GeneXus USA & Abstracta to stress test one of its internal systems that was undergoing infrastructure changes.

Our Solution

Abstracta conducted stress tests on the system, focusing on identifying performance deficiencies in the application. We installed the test software, created test scripts, executed test scenarios and analyzed the test results for changes.

Tools

Load testing: OpenSTA

Monitoring tools: Windows Performance Monitor, IBM iSeries Navigator, JMX and DbMon

Results

During the stress test, we implemented changes and improvements to the application.

We were able to prove that, using the new infrastructure, the application could successfully handle the peak loads experienced during the weekly “rush hours”.

We then reconfigured the new equipment so as to obtain the optimal possible performance from the application.

We improved critical transaction response times in the three main test cases.

We managed to reduce the time to complete the product shipping process from 763 milliseconds to 432 milliseconds at 50% of the estimated maximum traffic load.

Across the board, test case response times were reduced by 46% on average.

Furthermore, our stress test results provided documentation and recommendations for additional application performance enhancements.