When most people think of Docker, they think of microservices, 12 factor applications, cloud native architecture, and all things new. When Docker became a household name back in 2013, those were the architectures that they were targeting. Fast forward to DockerCon in 2016 and Docker, for the first time, stated that migrating legacy workloads was now their number one use case.
At DockerCon 2017 in Austin Texas this April, Docker announced a new service offering called MTA (modernize traditional applications). This offering was a combination of Docker technology, application migration methodology, and channel partners. Docker introduced Accenture as their MTA partner. Since May, Docker has added many other partners including Hewlett Packard Enterprise (HPE) and announced IBM as a new partner at DockerCon EU this Wednesday.
There are two hours a day of keynotes. On day 1, half of the time was spent talking about MTA and bringing up customers like Metlife to talk about their legacy application migration success stories. The other half of the keynote was to discuss Docker’s new support of container orchestration engine Kubernetes, which I’ll discuss later. The second day keynote focused 100% on MTA, to the dismay of many in the audience. In total, 75% of the keynotes were targeting legacy application migration.
I watched the complaints of Twitter range from boredom to “they are alienating young developers” which I got a kick out of. I don’t think young developers write big checks to Docker, but large enterprises who need to move 10,000 applications to the cloud in 18 months and shut down two thirds of their datacenters do. Sometimes boring pays a lot of bills. I also think this new focus is a reflection of the strategy of the new CEO Steve Singh, who was brought in to start driving up revenue.
Previous DockerCons were loaded with exciting new innovative announcements like great new security features, the Moby Project, Linux Kit, Docker Datacenter, and much more. This conference had two announcements and two themes. The announcements were “IBM is now an MTA partner and Kubernetes is now supported.” The themes were “Docker is a platform company and Docker can get your legacy apps to the cloud faster.” Let’s discuss each of these.
Modernizing Legacy Applications
Docker has raised over $190M and has successfully created a powerful brand in the world of technology. But making money in an open source model is not easy. Each DockerCon, Docker reveals amazing statistics that show their adoption growth.
But this is not the social media category. You don’t get valuated on eyeballs. You get valuated on things like revenue, profit margin, and EBITDA. Just like AWS, Microsoft, and Google, Docker needs lots of workloads running their software to make money. Docker is betting on two things:
- Most F500 companies want a hybrid strategy to avoid cloud lock in and need a solution like containers to accomplish that goal.
- Most F500 companies want to move many legacy workloads to the cloud.
Although microservices are the cool kid in town, there is a lot more demand in legacy enterprises to move legacy workloads than there is to build net new cloud native apps. Docker has acknowledged this and has doubled down on their investments and partnerships. They are so serious about this strategy, that they bored us to tears for two days pounding the drum on this message.
Think about it. What makes more money? A few new cloud native apps a year or moving 30 years worth of legacy applications to the cloud? The latter. Not many engineers like to hear that but CFOs like revenue and at this stage of cloud maturity, for most organizations, that is where the gold is.
Two years ago at DockerCon in San Francisco, Docker made some very significant announcements about their orchestration engine, called Swarm, which put Docker and Google in a head on collision course. As much as Docker fans love to stay true to the Docker tools and services, Swarm is still in its early days as far as maturity and scalability go. Kubernetes is the clear leader today for many companies leveraging containers. However, there was a big trade off if you did not choose Swarm. Swarm is deeply integrated with the entire Docker platform, especially in the area of security. Going to Kubernetes meant giving up some of the out-of-the-box integration causing the customer to have to do the integration themselves.
With this latest announcement, customers can choose either of Kubernetes and/or Swarm and get the same integration on Kubernetes out-of-the-box that they would get with Swarm. This is a big win for customers. In the analyst briefing, we pressed Docker on what it meant for the future of Swarm. Would they drop support of it eventually and concede to Kubernetes? The answer was no. Docker is going to continue to innovate on Swarm. The Swarm and Kubernetes communities are very close and a lot of the same people are on both. As Docker innovates on Swarm, if Kubernetes wants to embrace the new functionality then they will. But Docker does not want to be dependent on the Kubernetes committee because Docker has an aggressive roadmap and vision for orchestration.
Many people view this announcement as a surrender flag by Docker on orchestration engines. It definitely is an acknowledgement that Kubernetes is a market leader, but Docker will continue to make investments in Swarm.
Platform, Platform, Platform
Docker is pounding home the message that they are not a container company; they are a platform company. Containers may be the foundation of that platform but at the end of the day they are building the next generation of PaaS that allows customers to pick and choose components to create their own PaaS abstraction, without being prescriptive like traditional PaaS companies.
What Docker CTO Solomon Hykes told us is that they already tried PaaS and failed (referring to the company’s early days as DotCloud). This time they started over and built the platform from the ground up. It starts first with containers, but the endgame is the platform. Docker wants to be the central hub for both developers and operators for both building net new cloud apps and moving legacy workloads over to the cloud, both public and private.
Investors like platforms because they are sticky. Many F500s are planning hybrid cloud strategies to protect themselves from cloud lock-in. If Docker can be the platform of choice for many of those companies, Docker has a chance to build an enterprise platform play much like VMWare did many years ago. Now they need to execute and change their messaging from containers to platforms. Platforms is where the revenue is at.
The conference was not as exciting as the other four DockerCons I have attended, but it is probably the first time I left a DockerCon and had an idea of how they planned on hitting revenue targets to justify the $1B+ dollar valuation they have. Whether application migration is the right strategy or not will be determined over the next few years.
And while Kubernetes fans are celebrating victory with the Docker announcement of the Kubernetes integration, orchestration engines just became a commodity as I predicted a few months back. The money is not in orchestration, it is in platforms. As much as every company says they want hybrid clouds, I still see AWS as the platform of choice by a large margin. Only time will tell.