Today CoreOS announced the App Container specification with Rocket as a prototype Linux container solution. They are now requesting feedback and collaboration on the Rocket and App Container specifications.
Pivotal has put a lot of time and effort into managing Linux containers at scale with enterprise workloads running under Cloud Foundry. We have great interest in a composable, portable, open standard for Linux container image formats and runtime environments.
CoreOS’s etcd provides a key service for Cloud Foundry (and Kubernetes for that matter). Pivotal’s reliance on etcd established a foundation for a great relationship with CoreOS engineering, so we were intrigued when they approached us to collaborate on an open specification for Linux containers. When we saw the progress that CoreOS had made, and their openness to input and contribution, we decided that Pivotal needed to get involved in the App Container effort.
CoreOS has primarily become an operating system for deploying Linux containers and they’ve acquired considerable experience managing and supporting these systems. Pivotal’s container management has always been focused on the process lifecycle and telemetry with our emphasis on ongoing operations to fulfill the promises we make to our customers. CoreOS’ specification and engineering approach aligns with our perspectives and process management needs for running thousands of containers in production.
Pivotal believes in openness. We believe open source software is one of the most powerful forces shaping the ways modern industries create value. We steward a number of broadly adopted open source projects, some of which help power the most ambitious infrastructures and applications in the world. From Redis, behind the caches at Twitter, to RabbitMQ at the heart of OpenStack, to the popular and powerful Spring, and last but not least, to Cloud Foundry, Pivotal-sponsored open source software creates immense value across the industry for small startups to global enterprises.
Openness and open source is a spectrum. Open source with a code in a tar file is not the same as open source with open development and open intentions. Pivotal’s projects have their own history and momentum, each sitting at different places on that spectrum. Cloud Foundry has evolved from code periodically updated on github with little outside collaboration in the early days at VMware, to a more open process that includes published public backlogs, monthly community advisory meetings, and a soon-to-exist foundation. We believe this allows Cloud Foundry to create more value and would like to see a similar evolution for standardizing the fundamental Linux container building blocks.
The enthusiasm and appetite for Linux containers may seem new, but the features that enable containers were added to the mainline Linux kernel years ago. The Cloud Foundry container technology was based on kernel-level isolation application deployments from the very beginning. The development of that container technology has also corresponded with solving distributed scheduling, health checks, events, routing, metrics, and the rest of the details necessary to deliver a platform. Now that the industry is aware of and interested in Linux containers, Pivotal calls on the industry to create the POSIX equivalent for standardized containers that everyone can utilize with certain expectations and guarantees. Decoupling an open specification from the implementations decentralizes design decisions and allows the industry to explore the tradeoffs of those choices in more rapid innovation cycles.
Pivotal is enthusiastic about the development of open container specifications and their potential, and calls on the industry to review the CoreOS specifications, and get involved.
About the Author