I’m at DockerCon Europe today, participating in a panel discussion on orchestration. It’s an important subject for us, since orchestration is critical for our customers as they move from experimenting with containers to running containerized applications in production. Through our experiences with Distributed Resource Scheduling (DRS) on vSphere and vRealize Automation, we’ve grown familiar with the challenges of orchestration here at VMware. But containers present new challenges, especially around scale. To that end, several orchestration solutions have been developed by various players in the container space. We want to ensure that these new solutions are as well-integrated into VMware platforms as DRS and vRealize Automation were into our core vSphere technology.
Our goal is to simplify delivery of containerized applications to vSphere environments, so I’m excited to announce new integrations in the container orchestration space with Docker, Google, Mesosphere, and Pivotal. Our focus is on providing a common platform for building, operating, and managing applications at scale, and these integrations help our customers do exactly that. We’ll have more posts to talk about these integrations in greater depth, but I’ll provide some background here.
Docker has captured the industry’s imagination and customer interest around Linux container technologies, and we announced a partnership with Docker at VMworld earlier this year. Thus far, Docker’s focus has largely been on single host management, but its ambitions are to enable remote management and orchestration of many Docker hosts. Given that their orchestration technology is still in the early alpha phase, we wanted to focus on their upcoming remote management capabilities, called Docker Machine. Docker Machine enables developers and operators to start Dockerized applications on remote hosts, and today we’re introducing Docker Machine integrations for VMware Fusion, VMware vSphere, and VMware vCloud Air. This simplifies the process for deploying applications in VMware environments, whether it’s on a dev box via Fusion or into staging or production via vSphere or vCloud Air.
In June, Google introduced a container orchestration and scheduling system called Kubernetes, which has garnered industry and community interest. We started working with Google earlier this year, and today we’re introducing a way to quickly and easily provision a Kubernetes cluster onto vSphere infrastructure through a tool we developed called Big Data Extensions (BDE). As its name suggests, BDE was developed with a focus on big data workloads such as Hadoop. However, the types of problems we face when provisioning Hadoop frameworks are similar to what we see with container cluster schedulers like Kubernetes, so we used and extended what was already working for many customers. We encourage you to give it a try yourself!
We developed a similar integration with Mesos, the open resource management framework driven by Mesosphere. Mesos has an innovative two-level design that supports many different workload types using the same underlying infrastructure, and is already present in large production implementations, most notably at Twitter and Airbnb. Given the volume of customer interest in Mesos, we wanted to streamline the deployment of Mesos to vSphere infrastructure. As with Kubernetes, we’re leveraging BDE to simplify the provisioning process.
VMware actually had its own technology in the orchestration space several years ago – CloudFoundry, an open-source PaaS – which we spun out when we created Pivotal. Pivotal now has its own version of CloudFoundry, called Pivotal CF (PCF). While PCF has container orchestration capabilities, they’ve been deeply enmeshed within PCF. A new project from Pivotal called Diego will rewrite parts of Pivotal CloudFoundry in a more composable manner, allowing the Diego container orchestration layer to be separated from PCF. We’ll be working closely with Pivotal to enable the same tight integration into vSphere for PCF that we have with the other container orchestration engines listed above. In the meantime, Pivotal is joining us at DockerCon this week to demo .
All in all, we’re impressed with the rapid progress in the container space and excited about enabling our customers to more easily and seamlessly provision their container orchestration frameworks and containerized applications onto Fusion, vSphere, and vCloud Air. We’re looking forward to getting your feedback at @cloudnativeapps, and there’s more good stuff yet to come.
About the Author: Kit Colbert drives strategy and product development of third platform application solutions across the company. Previously, he was CTO of VMware’s End-User Computing business unit, Chief Architect and Principal Engineer for Workspace Portal, and the lead Management Architect for VMware vSphere Operations Suite. At the start of the career, he was the technical lead behind the creation, development, and delivery of the vMotion and Storage vMotion features in vSphere. Kit holds a ScB in Computer Science from Brown University and is recognized as a thought-leader on third platform, end-user computing, and cloud management trends. He speaks regularly at industry conferences, on the main stage at VMworld, and is the Cloud-Native and EUC voice for the VMware Office of the CTO Blog.