Case Study 1
Orange (Now EE) had no method of measuring the uptime and response time of its Citrix applications in production. It therefore needed a solution to monitor the performance and availability of these applications.
AC created scripts for Orange’s installation of Mercury Business Process Monitor (formerly Topaz Business Process Monitor); a tool designed to monitor the performance of production applications in real time. The applications to be monitored included one using legacy, terminal-based technology and one using Citrix. AC created scripts for both. In creating the Citrix scripts, AC used its Citrix Function Library; a library of functions for use with Mercury Business Process Monitor and Mercury LoadRunner. AC developed the Citrix Function Library to facilitate the creation of reliable scripts for testing or monitoring Citrix. In certain circumstances, Citrix scripts can become unreliable, owing to the graphical nature of Citrix and the need to handle bitmaps on the screen. The Citrix Function Library addresses these issues of unreliability and facilitates the more rapid creation of reliable Citrix scripts.
By introducing automation into Orange’s software lifecycle, AC provided a solution to measure the performance of Citrix applications in production and create an alert in the event of an application failure. AC’s top-down solution provides a clear end-user view of production monitoring and clear matrices on application availability and performance. AC has reduced Citrix application downtime at Orange and therefore provided its client a substantial cost saving.
Case Study 2
In order to test performance, T-Mobile required a simulation to be developed, of a 9 million customer load, spread across the front-end, batch processing and internet sub-systems.
AC completed the automated performance testing as per T-Mobile’s specifications. This included testing T-Mobile’s new mobile phone billing system. This system comprises a call centre management front-end, a billing engine and a facility for customers to view personal account data over the internet. Our team used a combination of Mercury Interactive LoadRunner software and Unix Shell Scripts to simulate the load of 9 million customers.
Our recommendations to senior executives formed a key part of their evaluation of the suitability of the new billing system for T-Mobile’s core activities.