Migrating Applications to the Cloud
Case Studies
Overview
For 43 years, SAS has helped organizations worldwide capitalize on their information and drive business value by leveraging data to make better decisions faster. Not only are they a leader in business analytics software and services, they are the largest independent business intelligence vendor in a market where many have been acquired by larger enterprises. Today customers at more than 70,000 sites rely on SAS to explore, analyze and visualize data, and transform that information into intelligence. To ensure that their software always performs as expected and delivers the vital insights their customers require, they selected Eggplant for their automated GUI testing needs ten years ago. During our decade-long partnership, Eggplant has been instrumental in helping them become more agile, modernizing their infrastructure in response to technological changes. A key component of SAS’ modernization efforts focused on rearchitecting the SAS service layer to be successful in today’s modern cloud-centric environments. As SAS customers move their businesses to the cloud, SAS is moving there with them. Customers have an increased expectation for SAS applications to be flexible and robust and scale appropriately to spikes in demand—something that was not possible with their legacy architecture. In today’s age of continuous everything, even large, established organizations such as SAS are facing new challenges in meeting demands. But by leveraging their micro services architectural platform and relying on Eggplant to test their technology, they’re giving their customers the power to deploy software in any type of environment and drive greater value from their SAS implementations.
Greater flexibility and control
Historically, the SAS Web Application Server was a monolithic application consisting of asset of services deployed as a single unit. In this environment, as new solutions and APIs are developed over time the size of the monolithic application grows—making it difficult to maintain and introduce change. In addition, their legacy approach presented challenges from a performance and productivity standpoint. If a patch needed to be applied to a service, for example, the entire application had to be stopped and restarted. And because services contained within the application were so tightly coupled the application was a single point of failure, meaning that an issue with one service could bring down the entire application. To address these issues and allow for greater flexibility and control, SAS moved to micro services and launched the newly architected SAS Viya® in 2016. The move reduced the possibility of a single point of failure, ensuring that the performance of other services is not impacted in the event one service is down. Most critically, embracing micro services keeps SAS relevant to their customers and gives them the technology they need to meet today’s continuous delivery expectations. The enhanced scalability and flexibility of SAS Viya gives customers the ability to use as much computing power as they need, when they need it, and where they need it-on a private cloud or on public clouds. In addition, they can partner with the major cloud providers to advance SAS Viya to be more cloud-native, embedding containerized SAS technology and integrating analytics around open-source software.
Immediate Results
“Within six months of using Eggplant, we automated 75% of our regression testing,” said Hodgson. “In addition to reducing costs, this has helped us quickly achieve quality improvements by being able to increase the number of times tests were run on each development cycle. We anticipate that we will be running at least 2.5 times the number of test executions by the end of this financial year.” Increasing the number of times regression tests are run during a development cycle has a direct impact on improving quality by ensuring more issues are caught when changes and updates are made to an application in development. “Regression test automation will also lead to more flexibility in our delivery, and ultimately will help us significantly reduce our development cycle time,” added Hodgson.
Return on investment
“Automation has been win-win for everyone involved in the development process,” said Hodgson. “The improvements in quality lead to productivity improvements because, by finding defects sooner, we are helping take the pressure off the development teams, allowing them to focus on the next set of changes, rather than the ones they did last month.” Rather than using the efficiencies gained as a reason to cut back on staff, the increased productivity of BT’s testing community has freed up resources to spend more time investigating why tests might fail and increasing testing scope and coverage. “We are already starting to see the benefits of a cultural change initiated by test automation, where our testing community is moving away from the more mundane, repetitive testing activities to higher-value input into the delivery cycle,” said Hodgson. “And with the cost savings we’ve realized through automation, we’ve basically been able to achieve these benefits for free.”