All About the DC/OS Open Source Project


This article is sponsored by Mesosphere as a Diamond-level sponsor of MesosCon North America.

In April Mesosphere, along with 60 partners (including Accenture, Autodesk, Cisco, Hewlett-Packard, Yelp and Microsoft), announced the DC/OS project, what has been called the first open and comprehensive platform for building, running and scaling modern enterprise applications. Ahead of MesosCon in Denver next week, we got the chance to catch up with Keith Chambers, product manager at Mesosphere, to learn more about this important open source project that has Apache Mesos at its core.

Keith Chambers, product manager at Mesosphere  What is DC/OS?

Keith Chambers:  The DC/OS project ( is the open source version of our popular Datacenter Operating System technology, which is the simplest way to build, run, scale and manage modern enterprise applications. When we say modern enterprise applications, we’re talking about applications that utilize technologies such as containers, microservices, real-time data processing, distributed databases, and more. Our goal is to make DC/OS the datacenter-scale equivalent of Android, in that it has the potential to democratize the development of application architectures and operational techniques previously reserved to companies such as Google and Facebook.

The DC/OS project is a software platform that’s comprised entirely of open source technologies. It includes some existing technologies like Apache Mesos and Marathon, which were always open source, but also includes newer proprietary components developed by Mesosphere that we’ve donated to the community and which are fully open sourced under an Apache 2.0 license. Features include easy install of DC/OS itself (including all the components), plus push-button, app-store-like installation of complex distributed systems (including Apache Spark, Apache Kafka, Apache Cassandra and more) via our Universe “distributed services app store”. We’re also tightly integrating our popular  Marathon container-orchestration technology right into DC/OS, as the default method for managing Docker containers and other long-running services (including traditional non-containerized web applications, as well stateful services such as databases). How is it an operating system?

Chambers: We call DC/OS an operating system for a number of reasons, including how users go about installing the things they want to run on it. DC/OS abstracts a datacenter full of servers into a single logical computer (i.e., 1,000 dual-core servers become 1 computer with 2,000 cores), which means developers and operators don’t need to worry about individual servers or VMs and can simply tell DC/OS about their task’s or service’s resource requirements. DC/OS is very versatile and can manage a wide diversity of workloads, ranging from Ruby scripts to Microservices in Docker containers to entire database systems. DC/OS makes it easy to run these workloads in an operationally efficient way, maximizing resource utilization and providing automated high-availability. For example. DC/OS will keep all tasks running even when there is a significant hardware failure by restarting failed workloads on different machines, and gracefully failing over stateful applications.  Why did Mesosphere open source DC/OS?

Chambers:  As a company, and as individuals, we believe in open source as the best way to drive innovation and adoption. We believe that DC/OS will create a revolution in the way developers build applications and the ways that organizations deploy them, at a scale and velocity that we’ve never seen before. We want to put the power of DC/OS into as many people’s hands as possible. The future of enterprise IT—whether running in the public cloud or a private datacenter—is going to look a lot different than what many of us have been used to. There will be new and exciting licensing models (i.e., open source rather than strictly proprietary) as well as exciting operational and architectural advancements. DC/OS provides an opportunity for savvy organizations to start on that journey today.

Organizations that adopt DC/OS will be more competitive in an increasingly software-driven world. DC/OS helps companies adopt the operational and architectural practices of companies such as Apple, Yelp and Netflix (and, at a broader level, Facebook and Google)—whose  infrastructure, and the applications that run on it, are directly tied to their corporate success—without having to reinvent the wheel. The large pioneering companies spent millions of dollars in R&D budget and deployed thousands of highly skilled engineers to build their internal systems, but that’s an unreasonable expectation for most companies. With DC/OS, organizations don’t have to piece together open source technologies or build their own homemade technologies. DC/OS brings this type of advanced datacenter environment to anyone.

We also wanted to encourage our partners and open source contributors to build datacenter services for the DC/OS app ecosystem. The DC/OS app ecosystem allows organizations to adopt and operationalize complex technologies such as Spark, Kafka and Cassandra in minutes, including automating some of the operational tasks and best practices. Open sourcing DC/OS introduces our partners to the broader community of users, and allows them to build applications and services on a platform that is not locked to a single vendor or cloud.

And finally an open source foundation means real portability for workloads between clouds, racks and hybrids. This is dependable OSS. If organizations have to create their own stack and get it running in multiple environments, they’re wasting time they want to be spending on their apps (differentiating their business!). More time to write the apps, less time wasted making the datacenter work.  What do you think will be hot at MesosCon this year?

Chambers:  You’re going to see a much broader base of users, in part because of enabling technologies like DC/OS, but you’re also going notice that Apache Mesos is the definitive kernel for distributed systems resource management. Ben Hindman,co-creator of Mesos, is going to show us how far we’ve come towards realizing his original vision of dual-level scheduler powering an operating system for the datacenter, which can be traced all the way back to the Berkeley AMPlab where Mesos was first created.

Whether you’re talking about Mesos as part of the DC/OS and the huge traction it’s receiving as the first operating system for distributed systems, or one-off cluster management solutions built for individual application frameworks — Mesos has become the de facto standard for managing the underlying infrastructure. I think you’re also going to see some major end users on display talking about how Mesos and DC/OS help them operate containers in production.

This article was sponsored by Mesosphere, creators of the world’s first Datacenter Operating System (DCOS). Learn more at: