The two companies built Razor based on Puppet Labs’ enterprise message bus and node inventory technologies with the goal of enabling system administrators to manage infrastructure “at the scale and agility needed for the cloud.” Thanks to the open-sourcing, Puppet Labs claims that its active community can benefit from Razor’s modular architecture to create plug-ins for extending support to any operating system or boot sequence.
Razor is an open source solution, licensed under Apache 2.0. It is now generally available can be downloaded for free from the Puppet Labs website.
Prior to the EMC World launch of Razor, I had a briefing with Luke Kanies, founder and CEO of Puppet Labs, and Dan Hushon, Office of the CTO at EMC. Commenting on how Razor can automate the IT infrastructure life cycle, Kanies said:
“What’s important about this is it’s real-time automation software for inventory, which means you always have a completely updated picture of what your hardware looks like. And then you have the first hardware provisioning solution, with the responsiveness and productivity that’s required for the cloud, so as your needs change, you can very quickly and easily transition your infrastructure. We’ve always been able to do this with the virtual and cloud infrastructure, but now we can do it with the physical infrastructure, to move quickly and productively on the hardware. And now you can automate every phase of the IT infrastructure life cycle, from bare metal to deployed cloud applications with great automation tools that are cross-platform, open and model–based all the way down.”
When I asked whether you could call Razor “semi-virtualization” of on-premise hardware, Hushon commented:
“I think it’s really hyper-automation, for agility. Agility is really one of the key goals of virtualization. Agility has really been missing from the more physical world. It’s just as applicable to virtualization environments, but in this case we really have a project that is allowing for that agility of being much more vendor-agnostic. …(we can apply that agility) right out onto a physical infrastructure so that I can very much take a piece of physical hardware, a node of computer, and I can decide to start it up today with CentOS. And then tomorrow I sit here and say, what I really want to do is take that node that’s running CentOS and I want to load ESX on it, then I’m going to put that CentOS image back on top of the ESX so that I can get a couple more tenants on that infrastructure. And then when that tenant desire goes away, maybe I’ll change it back down to bare metal for some reason. And to be able to manage the process forward and backwards, doing provisioning without really having to put someone with hands on a keyboard the way we do it today.”
According to Puppet Labs and EMC, other Razor capabilities include:
- Dynamic Image Selection: Razor automatically selects the correct operating system image based on the auto-discovered, real-time inventory data, removing the need for manual intervention whenever there is a change in hardware configuration.
- Model-Based Provisioning: Razor models the desired state of each hardware node and its operating system, automatically tracks provisioning progress toward this state, and can even decide when to re-provision. This gives system administrators full control over a node’s boot sequence and a complete log of its life cycle.
- Open APIs and Plug-in Architecture: Razor’s RESTful open APIs give system administrators full programmatic control of the rules and models that govern operating system image selection and hardware provisioning.
- Metal-to-Cloud Application Lifecycle Management: With Razor, Puppet Labs is able to provide automation for every phase of the IT infrastructure life cycle, eliminating the headache of trying to integrate and maintain multiple siloed management tools.
In April 2012 at the OpenStack Design Summit, Puppet Labs launched an enterprise-level implementation of OpenStack, along with community partners Cisco Systems (NASDAQ: CSCO), RedHat (NYSE: RHT), Morphlabs and eNovance. In a briefing with Talkin' Cloud, Kanies noted, “We’ve built a set of reusable models that are able to build, deploy and manage the core capabilities of OpenStack. And these are built, again — collaboratively — with multiple companies, so we’re confident that this one solution is the best way to deploy and manage OpenStack.”
Readers can click here for the overview of the Puppet Labs open community.