Dynatrace Perform Day 2015
Every IT organisation is pressured to simultaneously respond more quickly to business needs and provide stable, secure and predictable IT service, according to Dynatrace CTO in APAC, Rafi Katanasho.
That is very difficult to do for most businesses that are entrenched in legacy, but those that succeed at it, become high performers like big Internet players Google, Facebook or Amazon Web Services (AWS), and are able to deliver astounding IT success stories like 30-percent more frequent deployments, 8000 times faster lead times and more.
But how do more established organisations, like banks, governments, manufacturing, telcos and more, deal with sprawling legacy mainframes for example?
How does a bank with mainframe at the background, have mobile applications in front, and transition from one to the other? How does it also do it while ensuring that the mobile app is useful, relevant and engaging?
Dynatrace’s Regional Director in ASEAN, Koh Eng Kiong had rightly pointed out, “It starts and ends with the customer experience, not the application.”
Merging two speeds of IT
The APAC CTO observed two speeds of IT, one being industrial IT, which is matured, stable and uses IT as enabler to address business needs.
“Then there is digital IT, where speed is king and failing fast is encouraged,” he described, adding that, “Organisations are trying to move to a new architecture and create an API layer to communicate with the digital world, and which indirectly makes backend IT go faster too.”
A typical app performance lifecycle would have involved an app development team, testing to ensure an app can scale and produces the right metrics, before going into the production environment.
“But as apps became more mobile, people complemented this with cloud-based testing, to ensure applications can scale across various locations,” Katanasho said, and gave the example of organisations that leverage internal testing and cloud-based testing with last-mile testing.
Last mile testing is a global network cloud testing capability that is hosted by Dynatrace.
These tests yield field results of a website’s response time, the frequency with which it was being accessed and so on. There are also infrastructure metrics it produces like usage of CPU, memory, I/O and so on.
Katanasho pointed out, “But with the way that applications look like today, a simple transaction having to go through much more complex architecture, how does one test in that environment? How does one be confident that the app will do alright in the real-world?”
One step in the right direction would be picking the right DevOps tools. The right tools to support or even fully automate software development, testing and delivery processes, empower everyone involved to be more efficient, effective and reduce time to market.
Significantly, DevOps tool chains exist to allow devs to build more quality into apps, and support them in establishing better feedback loops between devs and operations.
Development and operations on the same page
No doubt, the old way of testing, blackbox testing is not sustainable, especially in this day and age of instant gratification and rampant social media usage. According to AWS figures, just one second of delay in website or app response time, costs businesses USD1.6 million annually.
For Dynatrace, the key is to light up every infrastructure, to see every single transaction and be able to follow and appraise every single piece of code.
Dynatrace plug-ins allow developers to see every build of their application, from very high levels all the way to the line of code. “Views are embedded within testing tools and these view are not just high level but are also actionable, and can be communicated to remote teams,” said Katanasho.
This APM solutions company boasts the broadest capabilities across the data centre, for every app, with ability to zoom in and zoom out. “In some context, you need both,” the CTO said.
This kind of deep dive into the code and network no doubt is intensive and requires some level of automation.
Katanasho also pointed out in conclusion how important and integral testing is in an organisation like Google. “Google does something like 3.8 billion automated tests every day in their IT environment, across all their applications. Think of it as a pipeline to smooth out the production line.”