The DockerCon user conference recently concluded in San Francisco, and the evidence of the container movement gaining momentum was clear — 2,000+ attendees, 70 sponsors and more than 40 exhibitors.
Add to this the approximately $500 million of VC investment in container-related companies to date; the lists of hot startups that are beginning to appear; the anointing of new paradigms by respected investors; and the user surveys with record-breaking buying-intention scores and there is no doubt that the container wave has propelled the DevOps engineer as a new power buyer in the enterprise.
I expect that the impact of containers will be similar to that of virtualization’s impact on the data center.
The key difference between virtualization and containers is well described by Mitchell Hashimoto, DevOps guru and founder of the data center automation and management company HashiCorp: “Whereas virtualization is like the sci-fi concept of terraforming a planet, embracing containers is more akin to building many isolated biodomes on a single planet.”
Painkillers sell better than vitamins.
Virtualization created one platform company in VMware, and it looks like Docker’s visionary approach has been rewarded with a first-mover advantage. However, just as how a host of other independent players who provided adjacent services were created in the virtualization era, I see greenfield potential for entrepreneurs founding container-related companies today.
Here are some of the opportunities in the container ecosystem:
Operating Systems: In a container world, you don’t need heavy operating systems, like traditional RedHat for Linux or Windows Server.
Several lightweight operating systems have been adapted or designed for containers, including RedHat’s Atom, CoreOS, Microsoft’s Nano, VMware’s Project Photon and Rancher Labs’ RancherOS. In addition, Rancher Labs provides a complete infrastructure platform for running Docker in production.
Orchestration: Google’s Kubernetes provides an orchestration and management solution for Docker containers in a clustered environment to accelerate Dev and simplify Ops. It’s being supported by CoreOS and others.
Mesosphere provides a data center operating system with support for containers and is an example of how to take multiple machines and make them look like a single computer. Mesosphere also optimizes the orchestration of Docker containers at scale.
Storage: To expand the use of containers for stateful enterprise applications, you’ll need a storage layer with enterprise features such as data persistence across nodes, container-level snapshots, clones, etc.
Portworx provides elastic scale-out block storage natively to Docker containers. It enables rapid deployment and scaling of both stateless and stateful enterprise applications.
Just like a host of storage providers carved out valuable areas in the virtualization era, including 3Par, Nutanix, Pure Storage and Veeam Software, I think containers will create a new breed of independent storage companies.
Networking: Similarly, there is a need for container-aware networking services. Weave creates a virtual network that connects Docker containers deployed across multiple hosts and enables their automatic discovery.
In a buy versus build decision, Docker acquired SocketPlane.io to provide offerings in this space. Similar to the services provided by F5 or Nicira in the software-defined infrastructure world, I believe there might be independent providers for similar services in the container world.
Security: Security is another problem to be solved for containers. Docker and a host of startups, including Twistlock, are providing offerings in this area.
Management: As containerized applications are deployed into production environments, there will be increasing need for management solutions, like an IBM Tivoli or HP OpenView in previous eras.
Examples of container-specific products and companies in this category include the monitoring service Sysdig, HashiCorp’s Atlas for application and infrastructure management and configuration management companies like Chef and Puppet, who have container-specific offerings now.
These are the primary areas ripe for innovation, and there might be others that break out as a large independent category. The business models are still in flux and monetization is lagging, as it typically does any time a new technology gets developed.
Throughout my career, I’ve seen that painkillers sell better than vitamins. Companies that solve the pain points will make money; companies that provide vitamins will have a harder time.
A good example in the container space is Docker itself, which was founded as a PAAS and whose pivot to become a container company is well-documented. Although it is still early times, examples of solutions that are tuck-ins versus stand-alone candidates have begun to emerge, with Docker acquiring Kitematic, Koality, Orchard and SocketPlane.
A couple of other examples of features might include companies that are providing PAAS for containers, development environments, UI interfaces and conversion tools that could end up being part of other platform companies.
The field is still nascent, and the recent Open Container news illustrated that coalitions are just being formed. The other wildcard in the container space is VMware, which understands the necessity of containers, but wants them to run within their virtual world.
However, it’s fair to say that VMware is being disrupted by containers today in the same way they disrupted the server hardware market over the past 15 years.
All these developments make it a very exciting phase for bold entrepreneurs who want to build giants of the containerized era, and fruitful times for investors looking to partner with them.