Digital transformation didn’t start with COVID-19. But migrating IT operations to the cloud has been ascendant since the pandemic necessitated digital-first and remote technology. Suddenly, it became essential that a world of employees could simultaneously access platforms in real-time and deploy new applications faster, without managing an on-site IT infrastructure. At the same time, customers wanted more.
But with more digital needs than ever, integrating business systems also got more complicated. AppDynamics General Manager Linda Tong sat down with John Furrier of TheCube and Garrick Linn from Match.com to discuss how Match.com uses AppDynamics full-stack observability to mitigate the added complexity of a post-pandemic set of growing services, apps and customer demands.
Watch TheCube session here or keep scrolling for some highlights.
John Furrier, TheCube host: How has the pandemic changed digital experience expectations?
Garrick Linn, Match.com: Match has been around for quite some time — we’ve been here for over 25 years — and a lot of the mentality that we try to bring is to innovate.
But the pandemic brought a lot of uncertainty. We weren’t sure how people were going to react. Was it going to be that everybody kind of hunkers down? Dating is something that requires human interaction on multiple levels. And it turned out that people were still very much interested in getting to a place where they could find human connections. And Match, as a premium product, tries to make that delightful.
And so we had our hands full, especially at the beginning, doing things like checking out video features: How does that work? What are the expectations? Is that going to creep people out? If we try to offer that, are they going to use it? How are they going to date? How are they going to talk? How can we make sure that they’re safe?
We have been using AppDynamics for years now, well before the pandemic. We continue to use it to gauge, not just the type of traffic and load, but also to look into new features and judge how they fit into this huge complex environment.
Some of those timelines that were more relaxed were very much accelerated.
Like a lot of companies, we had to figure out how to deliver on that.
JF: How has the pandemic affected your customers’ need for digital experiences?
Linda Tong, AppDynamics: More and more of these experiences rely on digital services and these amazing ways to connect with each other in a very digital space.
Expectations of customers have changed. You want applications to be simple, easy to use and delightful. But on top of that, you expect your application to be performant. You expect it to be secure. You expect there to be no hiccups whatsoever because now this is your way to connect with others. This is your way to find dates or go on dates. And the last thing you want is to watch your screen pixelate as you’re trying to have an important conversation.
And these kinds of experiences and these challenges, as people build more and more of these digital services to build these connections, frankly, require a lot more of folks like Garrick and his team. They now have to deliver amazing experiences with perfect performance, no security risks, no bumps in the night. And that’s really tough, right? Expectations have gone through the roof.
JF: What is your cloud journey, and what are your goals in your transition to the cloud?
Garrick Linn, Match.com: We’ve had our on-premises data centers for quite some time. We started putting our toes in [to the cloud], although it was kind of intense at the beginning, just trying to get people on board to believe it was possible.
We started out with a fairly small SWAT team managed, within a couple of months, working closely with our developers and security folks, to just demonstrate that we could do it. We managed to take something like 80% of our front-end traffic for most of the day, just kind of spinning that up, learning lessons from that, knowing what we didn’t know.
If we didn’t have AppDynamics, it would have been almost impossible to get a read on it. Here’s just one tidbit: We used to have a data center in Virginia. And so physics being what it is, there’s been a flight that we have had to contend with. For a couple of years, we hadn’t had the 30-millisecond or so round-trip latency. All of a sudden, we were back to the cloud that reintroduced this latency.
So what does that mean?
Will the apps glide by and absorb it? How do we track it? How can we figure out what the delta is between on-premises and the cloud? AppDynamics was what we used to get a read and see that it isn’t as good as we know we can make it, but it’s something. It’s a starting point. We can tell others why. We can show them the graphs. We can show them the data. We can convince them to do it.
We have focused this year on our affinity apps, a collection of applications we’ve managed to completely migrate over. We’re going to be running in hybrid mode for a while to be able to make performance comparisons.
JF: What does the future look like? How are you pushing the envelope?
Garrick Linn, Match.com: We’re a premium product, so we back that up. That means maintaining high availability.
Over the next few years, we’re going to be looking at:
- What have we already done?
- What can we move, in a piecemeal way, that makes sense?
- What are the things that we can rethink?
We’re also using AppDynamics as part of our containerization initiative. We’ve got lots of virtual infrastructure. But we’re asking, “What does it look like on-prem? In a container?” We go down the list of different things that might be different. Then we compare that to what it looks like in the cloud.
When we got into this, we didn’t think it was going to be done in six months. Even if we have to deliver those features at a much faster rate, we know that in the long haul, we have to make smart decisions and plan the capacity to get there.
JF: IT has shifted to the cloud, and the trend is toward a modern organization. What are you seeing as the complexity of environments for IT now?
Linda Tong, AppDynamics: The complexity that IT organizations are seeing now, as they fully adopt the cloud for all their new applications and start to migrate some of their existing applications over, is only increasing in complexity.
The way that you can virtualize your applications — break them out into millions of services, and the dependencies you have on third-party applications or SaaS services — add more data points to cover. You have to make sure that those things deliver on service-level agreements.
Wrapping your arms around that requires a partner to help you separate signal from noise because now you’re going into a world without simplicity. It has gotten to some point where it’s beyond what you can actually keep in your mind, beyond where you can just look at data and sift through and understand. You really need tools and systems that come together to represent your business to you in a new way and abstract away those layers of complexity. It’s not a full migration to the cloud.
There are going to be some applications that stay on-premises, in their traditional environments. And then some of them are going to go hybrid. Some will keep parts of the applications on-premises, and they’re going to start to modularize components. And so it’s not going to be a mass-scale migration that leads us to the promised land and we’ll deal with the cloud complexity.
It’s going to be ever-increasing complexity. As we now introduce so many variants of applications, so many variants of technology, what people are going to need is someone who can help them cover that entire estate and understand it at scale.
JF: There isn’t a tech stack to rule them all because you have different use cases and capabilities. Integration is huge. How does AppDynamics help you with this for Match.com?
Garrick Linn, Match.com: So we have quite a bit of infrastructure. A lot of it is shared, maintaining sandboxes for user data and that sort of thing. Now, navigating that space is always interesting. So for instance, one of the new things that we have coming out is Star.com. It’s out there right now. It’s a dating site that’s geared towards single parents. It does share some of the infrastructure, but we’re realizing what that means, how it’s different — how our registration flow is different, how our subscription flow is different.
What are the things that DevOps are actively trying to improve on and rethink? That’s one of the things that we try to figure out. We’re asking whether an app is a good candidate to move over to the cloud sooner or later. Or does it need to bake a little bit more?
Having established those baselines with the shared infrastructure, having a pretty good understanding of how they react and how they work helps us tee up new initiatives and get in front of those needs in a more efficient way.
Leaders of cloud migrations aspire to have the insight to pinpoint user experiences and problems immediately and resolve them quickly — across networks, systems and apps. But as our business systems grow, full-stack observability gets more complex.
In the case of Match.com, some applications stay on-premises, and others are migrated at strategic times. Customer needs emerge and new apps are deployed quickly to keep up. Capacity planning is expanded to a growing suite of apps. The result? It’s easier to plan, scale and meet customers’ expectations.
While the pressure to deliver outstanding digital experiences mounted during the pandemic, teams have learned that adding to their stack doesn’t have to mean a loss of visibility or service. In fact, working in the cloud can mean teams have access to the best insights to create the best possible user experiences.