Open Adoption Software Interviews: Sysdig

Loris Degioanni is a veteran OAS entrepreneur, first with CACE Technologies—the commercial company behind an open source net analysis tool called Wireshark, which was acquired by Riverbed in 2010—and now with an infrastructure monitoring startup called Sysdig. In this interview, he shares his thoughts on working with the broader container and microservices community; finding the right balance between community features and commercial features; and the feeling that comes when you build a project that can outlast you and your company.

---

JAKE FLOMENBERG:Tell us about Sysdig and where the company is in its OAS evolution—Project, Product or Profit.

LORIS DEGIOANNI: Sysdig is the visibility and monitoring company for modern infrastructure, in particular for microservice and container-based infrastructures. The idea behind the company started when, after leaving my last job, I realized that architectures were changing and that the tools supporting these architectures had to change too, because some things were just not possible with old approach. For example, getting visibility into infrastructure with a highly fluid, orchestrated and heavily equipped interface was a very big challenge. It was a challenge to get the data out, and it was a challenge because making sense of the data was frustrating.

I started approaching this problem as a technology architecture perspective, and decided to release my work as open source. This came from my background, because my previous company, called CACE Technologies, was the commercial company behind Wireshark, an extremely popular open source analyzer. I witnessed very early on in the open source movement the power and benefits of having a community air out what you're doing.

So the first thing that we released within a few months of starting the company was an open source tool called Sysdig, which is oriented around the workflow of troubleshooting and finding issues in the modern container-based infrastructure. Sysdig started generating a nice community around it, which also became the base for the development of our commercial product. This community is very important for us because it not only allows us to speak to people that have a specific use case, but it also makes it possible for us to learn from this community. We can listen and we can use that to make a product that solves a real problem for them.

The open source version was released around mid-2014, and has been downloaded more than 1.5 million times. Our commercial product, for which we have around 300 customers, came about a year later, based on what we learned from the community. Essentially, we have a product offering where you can mix the open source Sysdig with the commercial Sysdig cloud, and they work well together to create a complete solution.

How does Sysdig think about making money in OAS, including which features to include in the open source and commercial versions of the product?

We are a slightly anomalous open source company because typically—especially in some areas, like databases—the monetization happens by providing either support or some kind of enterprise features to your open source product line, but the product is exactly the same. Sysdig it is slightly different because the core technology is the same on both the open source implementation and our commercial solution, but they’re complementary. The open source version is for in-depth single-machine troubleshooting, while our commercial solution is a slightly higher-level, distributed, end-to-end monitoring solution that collects the data from different machines, brings it to a centralized server, and then offers a comprehensive visibility on containerized and microservices infrastructure.

We did something similar with CACE Technologies, my previous company,  and Wireshark. That tends to work pretty well in the area of performance management because monitoring and troubleshooting are two complementary functionalities, and it's important to have both of them working together. The open source and commercial solution can essentially work together and coexist in an integrated way.

Why do you believe OAS is the right model for your company?

I think that it's the right model not only for Sysdig, but for anything that has to do with the enterprise. There are clear benefits for the consumer—I'm not saying purchaser because you’re not necessarily a purchaser, you’re a user of the product and potentially you become a commercial purchaser—which means that OAS companies can expect to see adoption from companies of all sizes. Not only does it tend to create standards and help avoid lock-in, but open source also makes it better for an organization, especially a big organization, to work with a startup like ours, because you decrease the risk of the product that you're using disappearing and you being, essentially, screwed.

This is very real for me because I spent many years of my life working on network analyzers with Wireshark. And now I'm not involved with that anymore, but the project is still going, it's still very strong. The fact that people change in the project doesn't mean that the project is not there or cannot survive.

Open source is also a great benefit for the makers of the technology—people like me and companies like Sysdig—because it helps us be much more effective at bringing great technology to our users. You can learn from your users in terms of testing contributions and enhancements of the product, and doing deployments scale. Without a vibrant community around your project it's much harder and much more difficult to bring excellent technology to the market.

Now that you have customers and are generating revenue, is there anything that you wish you had done differently when you first released the open source project?

In my experience, the a hardest thing with open source and open adoption software is finding the balance between which features you give to the community as open source and which features you make people pay for. You constantly second-guess it because essentially you have this trade-off between creating a strong and healthy community—grabbing land, if you will—and making money. You build a large community by making your open source and freely available solution as powerful and comprehensive as possible, but maybe cannibalizing a bit the sales of your commercial solution by doing that.

With Wireshark and with Sysdig, I'm always wondering if, “Did I strike the right balance?” And I'm never sure. It’s really challenging and really complicated.

Looking back on Wireshark, would you consider that a successful project? Is there anything you would have done differently?

Certainly for me as a person and for us as a company, Wireshark doing open source was a game-changer. I still sometimes think about how impactful the decision of doing open source was for my life and in my career. We managed to start the company with just a few people out of school and we brought it to profitability, and to successful acquisition, by just focusing on offering something great to the community. It was a really a fantastic experience and I consider it, personally from the company point of view, a great success.

I learned many things from that experience, and I try to practice some of them at Sysdig. We’re definitely seeing very fast growth at Sysdig, so this makes me even more bullish that this is a model that can actually work and can actually scale. It's a challenging model and it's not trivial to make it work, but I've managed to do it twice. With Sysdig, however, I want keep scaling into a company that has the opportunity to become a very big business.

How do you work with the other players in the container ecosystem, which is full of other open source projects and companies?

The container and microservices orchestration space is extremely vibrant nowadays. Docker has been super successful at bringing the concept of containers mainstream, while open source projects like Kubernetes and Mesos take the containers as building blocks and make them the fundamental pieces for microservices. It has also been interesting to see how a big company like Google is really adopting this model with Kubernetes and just focusing extremely heavily into building a great community. They're really doing impressive work and the Kubernetes community is really incredible.

We wanted from the beginning to be an important part in this community so, for example, we participate in CNCF, which is the foundation behind Kubernetes and other other important pieces of the microservices and cloud-native ecosystem. Two years ago, we started integrating Kubernetes support in both our open source and commercial solutions.

There's always this parallel between the open core and the commercial tools, and with both of them you need a very good solid integration with the community and with the ecosystem that you're operating inside. We've tried to grow in parallel and be very active members of the community from the very beginning and bet on the ecosystem from the very beginning

Do you have any advice for entrepreneurs thinking about doing something with OAS?

One thing that is very cool as an entrepreneur, with this model of building companies, is that your work survives both you and your company. I mentioned the before how Wireshark is still thriving, even though some of the people behind it, and definitely the company behind it, changed over time.

Sometimes, your company is not going to do well, or it will be acquired and somebody else will be calling the shots. But the beautiful thing when you approach things as a creator in the open adoption model is that things survive even if your company is not there anymore. Your project is there for everybody, for the community it's still there.

Maybe it's a bit of a romantic point of view, but I find it beautiful to sort of give life to something that’s able to go beyond you as a person and beyond you as a company, and become something even bigger.