In a fast-paced business environment, software companies must release new products in a timely fashion or risk being overtaken by competitors. However, new releases are not only about new functionalities, but also about ease of adoption. IT departments must be confident that new products can run on their existing hardware; and that they are compatible with mission-critical software already being used by employees.
Here is how Mindtree helped one of the world’s largest software companies develop an effective tool to assess hardware and software compatibility of their customers' IT environments through robust testing.
The challenge
The customer's flagship software product is used by a majority of the world's corporations. It releases an upgraded product every two to three years. However, it is a challenge to get large enterprises that have hundreds of thousands of personal computers (PCs) running previous editions to upgrade. Two major challenges faced by the customer's enterprise customers was:
- Lack of a mechanism to assess if the hardware on all the machines in the network will be able to support the latest product
- Lack of a mechanism to inventory software installed on all PCs to determine their compatibility with the new product
The customer therefore decided to build a tool that would help their customers assess if the machines in their network were capable of running the new version without disturbing existing software.
The customer was looking for a testing partner who could work with them to ensure a high-quality, scalable and accurate tool; and approached Mindtree for assistance.
Our solution
Mindtree worked closely with the customer to understand their specific needs. We determined that the customer wanted:
- A tool that could scan 100,000 machines in a given network in less than 72 hours
- Accuracy of 98% or greater
- A cost-effective testing framework that could be used to for subsequent releases
The first major challenge in this engagement was the simulation of 100,000 machines which the tool needed to scan. All the machines needed to be on a network so the tool could discover them and start scanning. It was not possible to procure 100,000 machines and add them to a test network, so we developed an innovative solution that allowed testing under realistic conditions, without an expensive testing set up.
Once the approach was finalized, we created a test framework using a combination of batch scripting and user interface automation. The framework had modules that performed setup activities, started the discovery and scanning and collated results at test end. This enabled us to execute overnight tests and also use the weekends for longer running tests.
As the final step, we analyzed all results, studied the application architecture and database and suggested tuning recommendationsthat would help the customer reach their goal of scanning 100,000 machines in less than 72 hours.
Client benefits
- Helped the customer support scanning of 100,000 machines where previously it was failing at 70,000 machines
- Reduced overall scan duration by 45.4%
- Exceeded customer expectations by enabling overnight test execution
