It’s a great time to be an application developer, and it just keeps getting better. Today, a consortium of companies led by Google announced the Cloud Native Computing Foundation (CNCF), and Mesosphere is a founding Platinum member. We have been involved in the foundation since its conception, and we plan to remain very involved in what could become an exceptionally important open source community.
The stated mission of the CNCF is to provide developers with the right set of tools for building, deploying and managing next-generation, “cloud-native” applications. Loosely defined, this type of application is “container-packaged, dynamically scheduled and microservices-oriented.” It’s only natural, then, that Google has contributed its Kubernetes container-management framework as a foundational technology around which the CNCF will build.
Mesosphere is a big fan of Kubernetes. We have supported the technology since Day One and, earlier this year, worked with Google on an enterprise-grade version of Kubernetes that runs on our Datacenter Operating System (DCOS). All of our work on Kubernetes has been upstreamed into the Kubernetes open source repo. We also have Kubernetes project committers and a dedicated team working on the technology.
At OSCON this year, we kicked off our new training program with a Kubernetes bootcamp. Up next, In October, is an all-day container-management training session at O’Reilly Velocity in New York.
We believe the CNCF will be a force multiplier for the Open Container Project (OCP), which was announced earlier this summer (also with Mesosphere as a founding member). The OCP plans to develop a standard packaging and runtime environment for application containers, which are really the building blocks of working with containerized applications. The CNCF has the arguably more-ambitious goal of standardizing how all those containers are orchestrated to function as a working, distributed, service-oriented system.
Life beyond containers
There has been a lot of hype around the container ecosystem, which tends to oversimplify what’s required to use them at scale with production workloads. When Mesosphere thinks about the next generation of containerized workloads, we think in broader terms than what can be (or should be) packaged inside a Docker container. We think about how millions of containers can be orchestrated across clusters of thousands of computers. We think about the Internet of Things and machine learning and big freakin’ databases. We think about enterprise security, hybrid cloud computing and a platform that supports all these with high degrees of automation.
That’s why we built the DCOS.
And that’s why at the same time we’re putting resources behind Kubernetes, we’re also putting them behind other DCOS services such as Kafka, Spark, Cassandra and more. It’s why we continue to build out the capabilities of our Datacenter Operating System platform, which is based on the same Apache Mesos technology that powers Twitter (along with other important companies and applications) and that’s already installed inside large customer accounts.
We recently published a case study from Autodesk, which provides a good example of what I’m talking about. The company has built an event-streaming system for its cloud platform, in which new containerized services are managed by a homemade Platform-as-a-Service layer called Ochothon — which is a version of Autodesk’s Ochopod framework built on top of our Marathon service. (Autodesk has also experimented with tooling for Kubernetes, as well, but opted to run its production service on Marathon.)
Alongside Ochothon, Autodesk streaming data pipeline is powered by Kafka, RabbitMQ, Redis and other services all running on Mesos. Containers are just one aspect of this entire system.
Powering the next generation of developers
At Mesosphere we see an explosion of new applications and distributed systems being built at datacenter scale. The recent announcement of our Mesosphere SDK and developer program highlights our commitment to helping users run — and build — whatever types of services or applications they need to grow their businesses. Early partners range from database vendors to storage vendors to full-on systems vendors such as Cisco. It’s easy enough to imagine developers using our SDK to build integrations with GPU technologies and deep learning libraries to bring AI workloads to the DCOS, or a newly funded startup building a next-generation BI product on top of it.
Essentially, developers bring the ideas, and we bring the distributed, highly available, dynamically managed infrastructure platform to host them all. We’ve already seen the power and flexibility of this model, even within the world of container management (and, broadly speaking, Platform as a Service) where Mesosphere natively supports Kubernetes and Marathon, which you can use out of the box, or you can choose to work like companies such as Apple, Twitter, HubSpot and Yelp that have all developed their own PaaS-like frameworks on top of Mesos.
Let’s work together on the future of infrastructure
We’re building our company around the idea that IT should be an enabler of new ideas. Business shouldn’t be boxed in by clouds, servers, operating systems or any other false boundaries dictating what can run, and where.
Our DCOS platform is the materialization of this belief. It turns entire datacenters (and/or clouds) full of physical or virtual machines into a single pool of elastic resources and runs the types of services that today’s cutting-edge and tomorrow’s everyday applications will require. Containerized microservice-based applications managed by Kubernetes will be a big part of that future.