Dynatrace brings AI smarts to APM as enterprises transform

Profile picture for user pwainewright By Phil Wainewright March 8, 2017
Application performance management vendor Dynatrace believes it can outgun competitors by adding AI smarts to its APM platform

Stopwatch on speed blur street background © BillionPhotos.com - Fotolia.com
Application performance management (APM) used to be one of those nerdy obsessions that IT people took care of so that the rest of us wouldn't have to think about it. Today, though, the rise of digital and mobile applications has raised its profile. It's still a bit nerdy, but it gets respect — because if those apps aren't performing as they should, then the business suffers.

A new breed of APM vendors has come to the fore to take on this management role. While there were incumbent vendors who had focused on traditional enterprise applications, they were not fast or nimble enough to meet the new demands of mobile and digital application operations, ceding this ground to the newcomers.

Of the three now established as the leading players, two have won most attention because of their Silicon Valley roots and venture funding — AppDynamics, recently acquired by Cisco for $3.7 billion on the eve of its IPO, and New Relic, which joined the New York Stock Exchange in its December 2014 IPO and is currently valued at $1.9 billion. [Disclosure: New Relic recently became a diginomica premier partner].

The third is Dynatrace, founded a decade ago in Linz, Austria, a year before the other two. It soon moved headquarters to the Boston area after winning the backing of East coast investor Bain Capital, and then in 2011 it was acquired by mainframe tools vendor Compuware. Two years ago. it was spun out to become an independent company again with funding from Compuware's private equity owner.

Birth of a new product

This checkered history has sheltered Dynatrace from the Silicon Valley limelight focused on its rivals, but hasn't deprived it of market opportunity. It's grown to $400 million in annual revenues, says its longstanding CEO John Van Siclen. That's some distance ahead of New Relic's projected $260 million this financial year and a similar but sightly lower figure for AppDynamics, extrapolating from its pre-IPO financial statements.

Not all of those revenues come from the company's latest product, which the company's founder and CTO Bernd Greifeneder first proposed four years ago to adapt to the software-defined infrastructure of the cloud. Monitoring at this scale and complexity would need an AI core just to keep up with all the dependencies in such a hyper-dynamic environment, he reasoned. Launched two and a half years ago as Dynatrace Ruxit, that SaaS product is now known simply as Dynatrace and has become the vendor's flagship growth engine.

From the beginning, Dynatrace's secret sauce has been that it continuously records activity, rather than taking a series of snapshots as other vendors' products do. This it can do because it measures activity as it goes in and out of the application, rather than having its instrumentation sit inside the core VM, where there are restrictions on the amount of data it can hold at any given time. This in turn provides a rich stream of data for the AI to analyze and learn from.

That AI learning comes into its own in the new product, whose preconfigured instrumentation is integrated with cloud orchestration APIs, so that it can dynamically adapt to changing workloads, says Van Siclen:

I'd say at this point we have the most advanced cloud, microservices, DevOps platform in the market.

That's important because today's buyers are looking for end-to-end visibility, he adds:

APM five years ago had the director of IT checkbox because it was on the ITSM checklist — like change management, configuration, capacity planning.

Now it's about the business outcomes. Are my users happy? Are we converting? If we're not converting, why aren't we? It's the high fidelity view into the health and function of the application, no matter how complex it is.

Customer experience

Van Siclen finds he can often identify a prospect for the Dynatrace solution simply based on the customer experience he encounters as a consumer. Five years ago, he was buying a product at a major US retailer and the credit card terminal had gone down. As the sales associate took care of swiping the card manually, she confided that these problems had been going on for a couple weeks. He immediately called up the Dynatrace sales rep for that account and told him it was time to close the deal:

I said, 'Now's the time to stop talking and sell them some software and take care of them because it's one of my favorite places to buy shoes and clothes and they can't even check me out.'

There are three scenarios where Dynatrace typically sells its APM offerings, he says.
The first is for use in an operations war room, for monitoring performance and reacting to issues as they arise. The AI element plays an important role here in simplifying the task and surfacing issues, he says:

We want to show people that there's another way to imagine the IT of the future, where you're not putting a team of people looking at glass screens all day, looking for red, yellow, green, and then trying to mobilize teams of people. You actually can have a virtual assistant buried in that's smarter, quicker, faster, and knows more than any group of humans and can change the way IT management actually works.

DevOps and business

The second is in a DevOps environment, where teams are monitoring the deployment of new code and any issues that arise. Here, Dynatrace can run synthetic tests to check the performance of code prior to deployment, as well as monitoring performance once it's live in production. Says Van Siclen:

We've had a number of customers go from four releases a year to hundreds or thousands of releases a year.

What we see now is that a lot of the senior people have figured out, if I'm going to be serious about digital transformation, I better have a fabric in place that allows me to do DevOps better, actually have visibility and manage my apps better, and then also give me business outcome-based views so I can prioritize better.

The final scenario is where business executives are monitoring revenue, conversion rates and other business metrics. That's what will drive the most take-up, he believes:

Those are the three scenarios. We figure that the business one is the one that will probably capture the exec's attention the fastest and actually be what drives adoption wider.

Increasingly, too, enterprises are starting to move ahead with transformation projects where workloads are moving into the cloud environment. Performance monitoring is becoming crucial to the success of those projects.

We see our customers say 'Okay, we're going to shift everything up to the cloud in the next 18 months.' I think it clicked and the organization said, 'Well, this has to happen now. We can't wait anymore, so do it.'

It's not like APM comes as a bolt on after the fact. It has to be embedded into those migration projects because it's so critical for them to ensure that, okay, I'm on-prem with customers. My response time, if I measure it, is okay. What's going to happen when I'm going to move 1,000 applications in the cloud? Has anybody thought about, is my user response time going to be the same or better as it was before?

As you embark on these transformation projects, you need to make sure this is in place.

Those projects are happening across many industries, he adds — banking and insurance, retail, hospitality, airlines and many others. Typically, they embrace new mobile and cloud applications alongside older on-premise applications, and APM has to monitor all of them to ensure the required business outcomes are being delivered.

My take

Dynatrace has a slightly different pedigree from its main rivals, which may give it a useful edge in a fast-expanding market. We'll be talking to some Dynatrace customers over the next month or two to find out more about its approach and how it works in practice.