If there’s one golden rule of digital transformation, it’s that developers must be able to develop new apps and features—quickly, simply, and securely. If there’s another golden rule of digital transformation, it’s that developers must be able to push that new code to production—quickly, simply, and securely. Without these capabilities, it’s less digital transformation and more digital lip service.
There are many paths that end up at this destination of developer productivity, but VMware Tanzu Application Service (TAS) might be the fastest and easiest. It’s a platform that’s helped hundreds of organizations revamp their software development and operations over the past several years. We recently sat down for a deep dive with two TAS users—Jonathan Regehr of smart device manufacturer Garmin, and Paul Pelafas of insurance provider Health Care Service Corporation (HCSC)—who explained what it meant for their organizations.
“It's so simple, and yet it's so life-changing, the idea that all a developer has to say is, ‘This is where TAS is at and this is my source code; run it on the cloud for me. I don't care how.'"
— Jonathan Regehr, software architect at Garmin
TAS is about empowering developers
When speaking about how his organization was sold on TAS, Garmin’s Regehr looks back several years, to a time when his boss had heard about how frequently Netflix was deploying code and wanted to push Garmin in that direction. As part of a company hackathon, Regehr and a colleague worked with TAS (then called Pivotal Cloud Foundry). After a day of coding, they had a running application supported by blue-green deployments, auto-scaling, and overall infrastructure automation.
“All that stuff showed up quickly; it was easy,” he recalls. “And we knew that was going to be our path to be able to deploy multiple times per day, like Netflix.
“We wanted to be able to move at the speed the business was asking for. And we needed to move from a 60-day deployment cycle in some cases to a 16-minute deployment cycle. That was our big business driver; we wanted that agility.”
Eight years later—and now well-versed in the intricacies of how TAS automates everything from container builds to security patches for operations teams—he’s still impressed with the ease of the ‘cf push’ experience for developers. “I just keep saying, ‘You can pry cf push out of my cold dead hands,” he jokes. “It's so simple, and yet it's so life-changing, the idea that all a developer has to say is, ‘This is where TAS is at and this is my source code; run it on the cloud for me. I don't care how.”’
In fact, Regehr says his team at Garmin has now implemented a GitOps workflow that makes it easier for developers to experiment with new ideas, as well as to push code onto the platform. They make a pull request, he reviews it to ensure there are no glaring security concerns, and “30 seconds later they log into TAS and they have a brand new environment, or sandbox, to test out their hypothesis.” If it’s successful, they can present a working prototype to their managers without ever going through a drawn-out provisioning process or spinning up zombie cloud instances.
“It used to be, ‘Are we ready? Do we have the the right capacity?’ Now, it’s like, ‘TAS has got this; don't worry about it.'"
— Paul Pelafas, digital delivery lead at HCSC
TAS is about empowering operations
For HCSC’s Pelafas, who works closely with his company’s platform operations team, it’s the platform-level automation of TAS that really stands out. “Our operators love it,” he says. “It just makes their lives easier. They can have a life outside of just watching these things and making sure things complete. There's so much automation and security built into it. It’s kind of a no-brainer, really.”
That level of comfort and trust is especially critical during open enrollment season, when lots of people are online choosing and updating their insurance policies. “It used to be, ‘Are we ready? Do we have the the right capacity?’” says Pelafas. “Now, it’s like, ‘TAS has got this; don't worry about it.'"
Garmin’s Regehr can relate. He recalls the old days of release cycles, when an engineer might spend an entire day “babysitting” a software release to ensure everything worked correctly—an especially tedious process on something like an e-commerce site with lots of active connections and sessions. Today, releases happen frequently, while those same engineers are off creating value somewhere else.
“A proper CI/CD pipeline sitting on top of a platform just makes that process so much better,” he says.
“You can move on a whim, based on who's going to give you the best deal. This is power.”
— Paul Pelafas, digital delivery lead at HCSC
TAS aligns with Kubernetes
It’s hard to talk about platforms today without talking about Kubernetes, and Regehr is adamant that platform operators need to understand Kubernetes and where it works best. That means choosing the right workloads to run directly on the infrastructure API. There’s also a steep learning curve for Kubernetes, so operations folks need to expose the technology judiciously. He’s interested in the upcoming TAS for Kubernetes offering (now in beta) because it fuses the developer empowerment and automation of TAS with the infrastructure flexibility and automation of Kubernetes.
“When I look at Kubernetes, there's different capabilities than you have with TAS, but there's so many more decisions you have to make,” he explains. “There seems to be a lot of ways you can shoot yourself in the foot as a developer. But to take TAS and put it on top of Kubernetes, that feels like the best of both worlds.”
Pelafas says he’s particularly excited about utilizing TAS on Kubernetes so as to maintain the existing developer and ops experiences, while also expanding on HCSC’s multi-cloud strategy now that HIPAA compliance (among other requirements) is easier to achieve. As he notes, because Kubernetes is portable and allows you to recreate a cluster environment anywhere using a standard blueprint, “You can move on a whim, based on who's going to give you the best deal. This is power.”
“With a watch, for example, the on-wrist experience has to be awesome. But the off-wrist experience is also super important. Having the agility to continuously iterate on software-based features . . . is a really big deal.”
— Jonathan Regehr, software architect at Garmin
Better platforms, better software, better business
Ultimately, though, any improvements around developer productivity, operational efficiency, and better tooling are about shipping software to improve the business. Regehr says Garmin is far beyond the point where it thinks it can simply ship a high-quality hardware product and walk away.
“With a watch, for example, the on-wrist experience has to be awesome,” he says. “But the off-wrist experience is also super important. Having the agility to continuously iterate on software-based features—like analyzing your routes and pace, or finding jogging partners in the same area—is a really big deal.”
While that might seem natural for a connected device maker like Garmin, product agility is also becoming a big deal in health insurance. Pelafas notes that TAS is winning over holdouts within HCSC because they recognize how quickly teams using it can get early versions of products (aka MVPs) out to start generating value and see how consumers react. This is critical for insurance providers because they’re facing increased competition from startups and large web companies, and because new business is going to come from individuals who never knew a pre-Internet world.
“The generations behind us are only going to know tablets, smartphones, and digital-first experiences,” says Pelafas. “They’re going to take smooth and always improving user experiences for granted. If you're not offering them, they're going to look somewhere else.”
