August 17, 2021
Ep. #8, Resilience & Chill with J. Paul Reed of Netflix
In episode 8 of Unintended Consequences, Heidi Waterhouse and Kim Harrison speak with J. Paul Reed of Netflix. They discuss resilience engin...
For companies at the cutting edge of enterprise software development, category creation can be a necessary way to differentiate themselves from existing solutions. But conjuring a completely new category out of nothing is a delicate business; you have to stay close enough to existing tools and terminology to give your audience something to compare it to, but not stay so close that your solution is swept into the fold of what’s already on the market.
In a recent episode of EnterpriseReady, William Morgan, CEO and co-founder of Buoyant and creator of Linkerd, shared first-hand experience with category creation, from developing an early microservices infrastructure at Twitter all the way through to Buoyant’s current challenges in navigating competitors. In this post, we’ll look at his experience with Linkerd as a case study in how teams can think about category creation when existing categories simply don’t fit.
William first recognized that there was a need to reframe conversations around about their technology when he and his co-founder struggled to explain Linkerd to engineers. Because there wasn’t an equivalent technology, everyone they spoke to came in with assumptions about what they could (or could not) do.
“We started out saying ‘This is a proxy that handles RPC calls, and Linkerd is an RPC proxy.’ We tried to describe that to people, and people would be like, ‘I already have a proxy.’
We’re like, ‘No. This is different.’ They’re like, ‘I don’t even use RPC. We use HTTP.’ And we say, ‘Under our model, HTTP is a subclass of RPC.’”
Because the Linkerd team spent so much time struggling to define their technology using existing terms that didn’t quite fit, they lost opportunities to capture the attention of potential users.
The first step in creating a new category and getting open source traction was creating the right terminology to define their category. William and his co-creator Oliver Gould decided to start calling their tech a “service mesh,” which had no meaning at all. “[It] was a blank space that we could write into. Then we could say, ‘Let me tell you about this thing called a ‘service mesh,’ and this is what it does,’” says William.
William notes that they didn’t try to own or trademark the term “service mesh,” because they were focused on having a way to talk about their technology and get people to use it. As a result, other organizations eventually started using the term to define themselves as well.
Now that Linkerd’s creators had the language they needed to explain what they could offer, the next challenge to build awareness of service mesh as a category. William shares that this next phase was “a whole lot of conferences and a whole lot of elbow grease.” There were a few key components to getting service mesh as a category off the ground:
Because Linkerd was an open source project, they were able to talk about it at open source conferences and invited people to contribute it. In the early days they relied on tying their technology to their roots at Finagle, Twitter’s open source RPC system, to help audiences understand what they did. This connection to an established company and a solid use case helped create credibility for an emerging solution that the market didn’t have a strong understand of initially. But as Linkerd gained traction, its story began to stand up on its own.
Another challenge was understanding what conferences were the right conferences to go to. Initially they started with microservices conferences, but found that the audience of architects at these conferences wasn’t quite the right fit. “They wanted to talk about their 18-month microservices roadmap, and they wanted to talk about CQRS versus event sourcing,” says William. “It was all interesting, but I was like ‘This isn’t really going anywhere at all.'”
They then came into the Kubernetes community, which was full of practitioners looking for immediate solutions: “It’s not like there’s anything that’s really Kubernetes-specific in Linkerd, or at least in that early version of Linkerd. But from the market perspective, this was the audience of people who actually needed Linkerd and they needed it in a short timeframe, as opposed to the 18 month microservices roadmap.”
Another step in establishing service mesh as part of the larger devtools landscape was gaining third-party recognition. For many teams, this might mean pitching to software analysts and journalists, or writing extensively about your new category (as LaunchDarkly did for feature flagging). In this case, William and Oliver submitted Linkerd to the Cloud Native Computing Foundation (CNCF), which is part of the Linux Foundation “Philosophically, we really aligned,” says William. “They’re very focused on the set of technology that allow you to build these cloud native applications. Everything that Linkerd was being used for, we didn’t really know that term beforehand. We were like, “Yes. This is exactly the things that that Linkerd can help with.” Linkerd’s acceptance by the CNCF put it in the same bracket as established technologies such as Kubernetes and Prometheus, which helped boost its credibility and extend awareness within the market.
Note: Listen to the whole episode for more from William on the nuances of open sourcing Linkerd and its impact on their business model.
The double-edged sword of success in category creation is that as the market accepts the existence of a category, competitors will inevitably start cropping up. William says that the emergence of service mesh solutions from large cloud providers like Google presented new challenges. “When Istio first appeared, it was very frightening. Because we were used to operating in a world where we had no competitors. We invented the service mesh term, Linkerd was the service mesh. It was getting a ton of visibility interaction. Then all of a sudden, the 800 pound gorilla comes into the room.”
But major players like Google and VMware developing their own service mesh projects has been a good thing for Linkerd adoption, because tech giants’ investments in service mesh have helped validate the space. As a result, the Buoyant team has been able to shift focus to talking about their own product philosophy compared to other solutions, rather than fighting to gain validation and to explain what they offer.
The evolution of the service mesh category is just one of the many topics that William covers in his interview on EnterpriseReady. To learn more about his journey from Twitter engineer to startup founder, including how Twitter made the transition from monolith to microservices, listen to the entire podcast episode.