Piston Computing (previous CloudAve coverage), the OpenStack ecosystem player founded by former NASA architect, yesterday announced the release of enterprise grade OpenStack distribution called pentOS (for Piston Enterprise OS). It is an enterprise grade cloud operating system meeting the security and governance needs of enterprises wanting to use OpenStack for their infrastructure services. With this release, Piston Computing officially opens doors after being in stealth mode. In this post, I will briefly discuss the features of pentOS and then explain why I think it is interesting.
pentOS
pentOS is an OpenStack distro that builds on top of the OpenStack codebase adding some proprietary features to meet the security and governance needs of the enterprise customers. With OpenStack fast maturing into a production ready code and with some large service providers tapping into OpenStack for their public cloud services, enterprises are already looking into OpenStack for their private or hybrid cloud strategy. pentOS is accelerating this trend by offering the exact solution these enterprises need.
pentOS
- is custom built to meet the regulatory needs of big data
- has CloudAudit specs built in
Some of the other interesting features of pentOS include
- As I hinted in my post on OpenStack deployment last week, pentOS makes the deployment of OpenStack very easy. One can easily roll out an OpenStack cloud in minutes
- They use what is called as Null-Tier architecture which makes private cloud scaling pretty seamless, as simple as adding one server at a time
- Not only the deployment is simple and seamless, the use of their Cloudkey concept ensures that the security meets the enterprise needs. Also, by using a custom Linux distro that is just enough to have a cloud running, pentOS is drastically reducing the surface area for potential attacks
- Uses a modular design to allow third party to build add ons or interoperate
- Since it is built on OpenStack, it can easily let enterprises to tap into any OpenStack based public cloud service.
However, it should be noted that pentOS is not Open Source like OpenStack. It takes the core OpenStack code and adds proprietary components to it. Even though the team is contributing some of the CloudAudit parts back to the OpenStack project, the source code of the rest of the components are not available. In short, Piston has taken an open source project and converted it into open core.
But, why is it interesting?
There are many reasons why I consider pentOS to be interesting but I am going to focus only on two aspects, OpenStack ecosystem and CloudAudit.
- Ecosystem, ecosystem, ecosystem: Whether it is proprietary or open source platform, the success or failure is determined by the strength of the ecosystem around it. The impact of the ecosystem becomes all the more important in the case of Open Source platforms. A more vibrant ecosystem increases the longevity of the platform drastically offering the users necessary confidence to invest in the platform. Regular readers of this blog know very well about the worries about Rackspace hijacking OpenStack project through the backdoor. Even though there was no evidence towards such a move, the controversy refused to go down completely. As more and more players like Piston Computing emerge from this ecosystem, the chances of Rackspace or, for that matter, any of the big companies involved in the project to manipulate the project goes down drastically. The more there are players like Piston Computing in the ecosystem, the greater the chances for market forces to ensure that OpenStack project will stay true to their initial stated mission.
- Transparency, transparency, transparency: As many in the pundit world (including me) were emphasizing, one of the key factors necessary for large scale cloud adoption is transparency. Unfortunately, most of the cloud services providers are either not being transparent or very slow to embrace the idea. A big part of the reason is the difficulties involved on the provider side. It is just not possible to open up their infrastructure to all and sundry to come in and poke around. To overcome this limitation, a group called CloudAudit (now under CSA) put forward specs to allow the consumers of cloud services to programmatically query the service providers to see if they meet their security needs. Companies like enStratus have already implemented the CloudAudit specs but none of the cloud infrastructure platforms had CloudAudit baked into it. With the release of pentOS, the situation changes and it appears that the Piston team is going to contribute their CloudAudit work back to the parent project. This will help more and more service providers to embrace CloudAudit giving the much needed transparency in the cloud ecosystem.
Conclusion
This is clearly an interesting release and greatly strengthens the OpenStack project. I am waiting for Nebula, the company formed by NASA CTO in the OpenStack ecosystem, and CloudScaling to come out in open about what they are doing with OpenStack. As more and more companies build their business around OpenStack, we are going to see the repeat of history, like what happened in the Linux world. OpenStack will be similar to the little kernel Linus Torvalds wrote that changed the world of IT upside down. Such an evolution will be good for the consumers of cloud services, empowering them to control their destiny as they use more and more cloud services. In this sense, pentOS is a welcome addition to the OpenStack ecosystem.
Related articles
- Piston Puts Enterprise Cloud On A Memory Stick (informationweek.com)
- Piston launches OpenStack cloud OS for private clouds (zdnet.com)
- Piston Cloud Launches pentOS, An Enterprise OpenStack Distribution (techcrunch.com)
- The Dream of OpenStack in a Commercial OS is Realized (readwriteweb.com)
I think you are missing the point with your ecosystem argument. The issue is not that one vendor (Rackspace or anyone else) is going to hijack openstack, the issue is that they are all going to do exactly what Piston has just done. Release their own proprietary version of openstack. This is possible because openstack is Apache licensed, so any vendor can take any or all of it and make it closed source.
What this will result in is numerous different flavors of openstack, none of which will interoperate with each other. This is where the ecosystem argument breaks down, all of these vendors will claim to be “based on openstack” but since they are all making their own additions and modifications to the code, none of these will interoperate with each other.
We won’t end up with a single “openstack ecosystem” but instead numerous competing vendor infrastrutures (the Citrix openstack infrastructure, the Dell openstack infrastructure… and one for Piston, and one for Nebula, and one for…)
So this grand vision of everyone uniting around openstack will instead lead to balkanization (division)… and if that’s true then what’s the point of openstack anyway? It’s gone from being a standard that unites to a marketing term that promises unity but doesn’t deliver (“We’re openstack based”) without every gaining any commercial traction.
Nick, the concern about Rackspace hijacking OpenStack was real till sometime back. There may be some players who completely fork OpenStack and make it proprietary but they can ONLY survive if they have the enormous resources needed to manage the proprietary fork. Almost most of the players are going to keep OpenStack at the core and play the Open Core game to tap into the community under OpenStack. As long as they keep OpenStack at the core, it is going to be interoperable. The chances of your fears coming true are very slim. How many proprietary forks do we have on Apache?