Build an Open Cloud – Getting Started with OpenShift Origin and OpenStack - Archived

What’s the difference between OpenStack and OpenShift Origin?

As I talk to people about OpenShift, and now about the OpenShift Origin project, they are sometimes confused about how it is related to OpenStack.

OpenStack and OpenShift Origin are different but complementary open source projects.

OpenShift Origin is not part of OpenStack, but it can run on top of an OpenStack system. Just like Apache and MySQL are not part of Linux, but can run on top of a running Linux distribution like Red Hat Enterprise Linux.

What is OpenStack?

OpenStack is a large open source project developing a cloud Infrastructure-as-a-Service, or IaaS. It can roughly be summarized as an “open source version of Amazon Web Services”. Several different companies are contributing to OpenStack, including Rackspace, Nebula, and HP Cloud. Red Hat has announced it is joining as a member of the new OpenStack Foundation.

OpenStack itself has a number of large sub-projects. Swift is an object store for containing bootable images, chucks of data, and web resources like HTML and graphics. “Nova” is a system for running and managing executing compute resources, which are virtualized operating system images.

What is OpenShift Origin?

OpenShift Origin is an open source project developing a Platform -as-a-Service. A PaaS orchestrates and automates as much as possible for the 95% of the use cases involving web applications written in well-known languages and frameworks that is using well known relational or NoSQL data stores, such as MySQL or MongoDB. For a detailed article on what a PaaS is, check out Dan’s blog post, “What is Platform a s a Service (PaaS)?”

IaaS vs PaaS

When you use a IaaS, such as OpenStack, you give up the raw efficiency of running on bare metal and of being able to tune your own hardware and storage to your precise needs. But by trading that away, you gain enormous flexibility and speed in scaling and size, in easily allocating only the resources needed, and reduction in CapEx risk. You can add and remove machines and storage and internet bandwidth in seconds, not in weeks or months.

When you use a PaaS, such as OpenShift, you give up the flexibility of having complete control of the orchestration and scaling of your application. But by trading that away, you gain enormous speed of deployment, and a reduction of developer-hours spend working on “plumbing” not directly related to your own application and your own customers.

IaaS and PaaS cloud techniques are not in competition. In fact, PaaS systems typically run on top of IaaS systems. For example, the existing hosted OpenShift PaaS operated by Red Hat is currently hosted on Amazon Web Services.

We expect that as people build public, hybrid, and private PaaS systems using OpenShift Origin, they will want to run them on OpenStack IaaS systems. These OpenStack systems may be operated by cloud providers, in internet-facing private datacenters, or even inside an intranet for enterprise private applications.

How can I try it right now?

Start with having VMware or VirtualBox installed on your machine.

Download an OpenShift Origin LiveCD image from /open-source

Look for instructions on how to deploy OpenShift on top of OpenStack instructions on our wiki in the coming days.

What’s next?

We expect that interest in OpenShift Origin on OpenStack will be high, and that it will be an area of a great deal of community development. As two open source projects working together, who knows what we will build together?

Comments are closed.