New version of Linux for the containerization of cloud servers is now available on a monthly subscription basis.
CoreOS, a new version of Linux designed for running containers on masses of cloud servers, is available for the first time as a supported subscription service.
CoreOS Managed Linux is CoreOS's supported version for running on servers in Amazon's EC2, Google Compute Engine, and Rackspace, among other cloud services. CoreOS is a minimal version of Linux with a kernel, associated utilities, and little else. It amounts to about 100 MBytes of code. It can boot in 1.5 seconds out of solid-state storage, and it's designed to run hundreds of containers on a single host.
"We're the equivalent of 25 pictures on your phone," quipped Alex Polvi, co-founder and CEO of CoreOS.
CoreOS is also designed to provision Docker containers, a Linux container format that has become popular with developers over the last 15 months. CoreOS's capabilities with Docker mean that it may emerge as a challenger to Atomic Server, Red Hat's coming version of Linux for running Docker containers in the cloud. CoreOS executives said a technical support subscription for a four-way Linux server will cost about $1,000 a year, or about one-fourth of Red Hat's public price list for support.
CoreOS Managed Linux will be a checkbox option as customers configure workloads on Amazon, Google, and Rackspace clouds. Rackspace also recently announced OnMetal dedicated servers, designed to run large enterprise applications in the public cloud. OnMetal servers will use CoreOS as their operating system.
In addition, CoreOS is open-source code that can be downloaded and used on-premises. CoreOS, the company, is based on a pay-for-support revenue model and doesn't have a proprietary offering with more functionality than the open-source code, Polvi told us in an interview.
"We can boot in one-and-a-half seconds, and our memory footprint is much smaller than traditional Linuxes," Polvi said. The fact that the company has containerization cred and is gaining traction in the cloud has not been lost on the venture capital community. The 15-employee company announced Monday that it has received its first funding from Kleiner Perkins Caufield & Byers, with follow-on investment from Sequoia Capital and Fuel Capital, for a total of $8 million.
"But Google Chrome trumped everybody. It just automatically updates itself. You're always running the latest version."
Server operating systems have been the opposite type of software. They're configured on a server and then left alone indefinitely. Patches are reluctantly applied due to the potential hazard of changing something that disrupts the applications running under them.
CoreOS automatically updates itself.
CoreOS in the cloud will be updated automatically. Even running production servers will be updated frequently and automatically, if their supervising administrators grant permission, using techniques designed to avoid interference with running workloads.
"It's the ideal setup for running your infrastructure. The smartest guys in the room are doing your maintenance. You're always running the latest version," says Polvi.
CoreOS can produce such a server system because it's learned the lessons of Google and Facebook, which run Linux containers on their cloud servers, and it's come out at a time when "there's a movement to this warehouse scale computing. We're very lucky to be part of the next generation."
In addition, CoreOS's staff has cloud engineering experience. Polvi is the co-founder and former CEO of Cloudkick, maker of a server management platform, acquired by Rackspace in 2010. After the acquisition, he headed Rackspace's Bay Area R&D effort until leaving to found CoreOS.
He is joined by Brandon Philips, former manager of the Linux kernel at Novell/Suse; Jonathan Boulle, engineer in cluster management at Twitter; Blake Mizerany, one of the first engineers hired at Heroku; and Brian Waldon, project team lead of Glance in the OpenStack project. (Glance manages virtual machine images in OpenStack.) Polvi says CoreOS's small staff is primarily made up of engineers with similar backgrounds.
CoreOS is considered by some observers to be a fork of Google's Chrome OS system, customized for Linux server management. The system is so small because container workloads contain part of the Linux operating system themselves, the user-space parts needed by the application. But all kernel functions, such as scheduling processes and memory management, are the function of the host system and shared by whatever number of containers is running on the host. Containers also leave each workload isolated from the others in a manner that's sometimes described as "lightweight virtualization."
CoreOS leaves out all the desktop and end-user oriented features of Linux in favor of being a minimal server system. Unlike Red Hat Enterprise Linux, Suse, and Ubuntu, it has no package manager scrutinizing open-source code modules for possible incompatibilities. Since each workload runs in its own container, it's up to Docker to enforce correct package management.
CoreOS will also run on service providers BrightBox Cloud, Vultr VPS, and VEXXHost Cloud. It will also run under OpenStack, Eucalyptus Systems, and other private-cloud configurations, according to information on the company website.
You're mistaken if you think you can execute in the cloud without using software to orchestrate application lifecycles. Here are four reasons you must. Get the new Cloud Automation Tech Digest today. (Free registration required.)
Charles Babcock is an editor-at-large for InformationWeek, having joined the publication in 2003. He is the former editor-in-chief of Digital News, former software editor of Computerworld and former technology editor of Interactive Week. He is a graduate of Syracuse ... View Full Bio
Multicloud Infrastructure & Application ManagementEnterprise cloud adoption has evolved to the point where hybrid public/private cloud designs and use of multiple providers is common. Who among us has mastered provisioning resources in different clouds; allocating the right resources to each application; assigning applications to the "best" cloud provider based on performance or reliability requirements.