There are literally millions of problems that can send your company’s new app off the rails and cost millions of dollars, but there’s one thing all those problems have in common: you probably won’t see them coming.
Why? Because unless your developers are testing their code – and, importantly, unless their machines are set up for the rigorous testing they should be doing – your company will be like the straight man in a comedy skit who never sees the cream pie before it smacks him in the face.
Or, like the hipster in the cartoon above, your test machine will be set up to look for one outcome and nothing else.
That’s really the essence of what service virtualization does for companies. It helps you anticipate and avoid the unthinkable. SV allows you to create a kind of software “wind tunnel” where you can twist and tweak the knobs to simulate any environment or condition without risking any harm.
The reality is that the majority of developers involved in Agile release cycles don’t do performance testing of any kind on their code before pushing it out to production. Unfortunately, the mantra is “just make it work,” not “let’s make it perform to expectations without crashing the whole network.”
This is not because developers are bad people. Most likely, organizations simply don’t have the time, resources, or budget to replicate production environments to test every Agile release. As a result, they end up testing in production outside of work hours. How big of a mistake can that be? Ask the folks at Knight Capital, who released a test into production and cost the Wall Street fund $764.3 million in losses when its trading program went haywire in 2012.
However, averting such a colossal mistake is only the first of many reasons your company should be using SV to test in life-like environments for any situation. Just as important: testing application performance and not leaving it up to your customers to decide your brand is not up to snuff.
As Kieran Taylor, an APM and DevOps guy at CA Technologies, told APM Digest recently, “the top factor impacting application performance is being able to clearly see the end-to-end picture of your application.”
“Having the ability to simplify and streamline the end-to-end view of your application lets you see just what you need to see to pinpoint bottlenecks before they impact users,” he says.
In today’s software-driven world, where all the devices around us are in constant conversations with other devices (read what Gartner has said about the Internet of Things), your developers face unprecedented integration challenges.
Take the example of Denmark’s Copenhagen Smart Cities project, an unprecedented experiment in linking everything in the city, from intelligent lighting to parking sensors to trash cans. Project directors’ key goal is to “de-risk the real-world trialing of vendor technology.” They’re using smart simulation on an urban scale, using digital engineering principles such as systems dynamics and systems thinking to test the city’s physical capabilities.
Smart simulation is necessary because we live in a messy, complicated world, with many unthinkable possibilities. In such a complex environment, a simulation built to test only one scenario is pointless and futile. The same is true for any app you want to deploy into a live environment.
Service virtualization can help your organization turn that corner by virtualizing all of the systems your app will encounter in the real world, whether it be a third-party signature pad, identity management system, payment processing app, or reservations app.
Remember, no app is an island. Without rigorous integration and performance testing, it’s doomed to sink.
If you’re interested in hearing me talk more about virtualizing IoT devices and other fringe use cases, I invite you to listen to the Tales in Testing - Paranormal Service Virtualization Use Cases webcast.