June 07, 2018

#007: Taking Care of No-Downtime Deployments and Monitoring

00:13Dirk: Hi everyone, welcome to a new episode of Commerce Tomorrow. We have an exciting technical episode today for our CIO and CTO’s listening. Welcome Olaf, CTO and co-founder of Vamp. Olaf, would you mind introducing yourself?

1:04 Olaf: Yes, thank you for this opportunity! My name is Olaf Molenveld. I’m located in Amsterdam. I have a technical background, working in the internet industry since 1995. I did a lot of programming in the past, and then moved to consultancy and later architecture. I’ve now jumped onto the cloud, microservices bandwagon. Initially, we started building an ecommerce platform called Magnetic.io, but it was a little too early at that point to be successful. We had some cool internal tech that we used to start Vamp, our current and only product that we are developing. I try to be in between tech and business by making sense of technology.

3:12 Dirk: Can you explain what Vamp stands for?

3:22 Olaf: It actually started out as an internal joke. It was a very awesome microservices platform inside Docker containers. Around 2014, when we started, the technology wasn’t really there so we started building this for ourselves. We started out on Mesos. People were very intrigued about dashboards that we were showing at meetups. They were asking to buy it before it was even for sale so we started thinking about making this into a product for other companies.

5:05 Dirk: That’s actually a really great story about how products get developed.

5:38 Olaf: Absolutely, this is resonating in the ecommerce space. They need to push the envelope forward. It really works if you can explain to big monolithic ecommerce stacks. You want to move faster without more issues, so you need more tooling around that stuff.

6:35 Kelly: That makes perfect sense. I saw a demo you did a few years ago and was pretty impressed. If you could explain, what are the problems you are trying to solve?

7:04 Olaf:If you step back and look at how typical releases are done, it’s often one big bang. You test internally and then flip the switch. If you compare that approach to retailers in the real world, everyone would call you crazy if you did this. This is how a lot of software releasing still happens. It’s because people don’t know any other way. When you introduce microservices into the equation, you are able to evolve faster. You also need something to package so you can deploy, which is Docker. We saw the possibility of Docker early on. You have everything packaged in Docker. You also need tooling to move quickly from version A to B without any downtime. This is where Vamp comes in. You are able to get things into production without any downtime. You have different components that you need to coordinate. We make it possible to run to the new version quickly. You are also able to aggregate metrics so you can make sense of how they perform.

12:00 Dirk: I think that is delivering great value for everyone shifting to microservices. What is the maturity level on technology and microservices that a customer coming to you needs to have?

12:45 Olaf: There’s a few dimensions: technical and organizational. The technical things that you need to have in place are a modern stack where you can run these containers and monitor. You need to be able to develop microservices where those services are still statically talking to each other. The more challenging things are in the organizational thinking. You need to move from a project drive organization to a more data driven organization where you try things out to see how they perform. This is the level where people realize that this is possible. This means that business needs to get back into the driver's seat.

17.04 Kelly: Retailers today aren’t innovating very well and I think your product helps to solve this. We see retailers releasing monthly (on average) and they need to embrace innovation. Iteration drives innovation. When you entered the market, Kubernetes was just a recent project. Clearly, there a lot of changes in the container orchestration space. How has your product changed over the last couple years and how are you evolving?

18:13 Olaf: Our vision has always been to provide baseless value and functionality. The moment Kubernetes came in, we moved up also. We are working on deeper integrations and try to make sense of all technical innovations. We are a control system for underlying technologies with drivers against these tools or serverless APIs. We provide functionalities so they can easily try things out.

12:10 Dirk: Talk a little bit about customers. What kind of customers do you see most often using your platform?

20:31 Olaf: We see a lot of ecommerce platforms since they’re moving quickly. We noticed that they often start moving into more hybrid architectures. There is an experimental mindset with retailers. They seem to be more responsive to our vision to deploy faster and at the same time measure. Financials were also interested but having a hard time moving from innovation departments to production. We do see a lot of traction from telecom and cable companies.

25:50 Kelly: The cloud native computing foundation named you to their cloud native ecosystem. If you could explain what cloud native computing foundation is, where they fit in the market, how you define cloud native and if you have any thoughts about where they placed you?

26:19 Olaf: To be honest, we are not official members. It’s a foundation of companies that make tools and software that is modern, build around schedulers and containers. It’s exploded over the last few months. They positioned us under service management, which makes sense. There are so many little areas going on there. In a while, it will die out and come more compact.

28:13Dirk: Let’s use this as a chance to talk more about entrepreneurship. How is it starting a company in the Netherlands? What are your growth plans?

28:43Olaf: Of course it’s great to be in Amsterdam. We attract a lot of talent here. There’s a strong work life balance. Our office is located in the city center, on purpose, to attract engineer talent. We of course have our challenges to find the talent. I think Amsterdam is very well positioned though: commute, travel, weather, international are all great reasons to be here. We are looking to branch out. Germany is a mature market on our list, Silicon Valley and the US also where they are generally moving faster than Europe. We are already global but building up a client success team in the US. Most of the support is here in Amsterdam.

32:00Dirk: It’s a great city, a great place to be. If you have the chance to go, you have to.

33:06 Kelly:One more question about entrepreneurship. How is your commercialization model?

33:22 Olaf: When we started out, we had a vision of open source, single tenant core. We thought this would help us sell the commercial product. That didn’t really work out. Vamp is also more of a platform with a philosophy than a technical point solution. It’s not crippled in any way on functionalities. We are a product company so we are fully focused on this and the professional services around it. We are working with partners that use our tool in their projects.

35:42 Dirk: Do you see most of the stuff being built in house or working with an agency?

36:09 Olaf: It’s a hybrid model. The companies that can build in house, do. Like Netflix, Spotify or Facebook. There are also the companies that don’t want to reinvent the wheel who are looking for development outside. In the end, we will end up with a hybrid model. You need to have a certain level of internal understanding but not necessarily the ability to build yourself.

38:38 Kelly: I have one final question for you. You’ve seen a lot of organizations attempt microservices. Can you give your thoughts on when someone should adopt microservices?

39:13 Olaf: That’s a very interesting question. You see it fail quite often. You need to run into the limitations of a single service application. The things is moving to microservices is more complex because the services need to talk to each other. It’s more complex than a single application. You need to have a strong internal business case of why you want to make this shift. Simply adopting microservices to have less issues is a no no. You need to understand how this will help you and why you need to make this architectural decision.

41:33 Kelly: That makes a lot of sense. Thanks so much for your time today!

41:40 Olaf: Thank you very much for the opportunity!

Discover and listen to our podcast through