When NASA decided to join forces with RackSpace to announce the start of an open source cloud project under the name of OpenStack it was a natural headline grabber. Anything with the tags ‘Clouds’ and ‘Open Source’ gets attention as working CIOs and IT departments grapple with what the technology industry tells them is their future. Add NASA into it and you have the hope that these guys really know what they are doing – there is a joke there somewhere about rocket science – and maybe its all going to make sense.
The lack of clarity and straightforward information sources is what drives my blogs on the topic; at the moment we are in the hype cycle with each major vendor trying to position their approach as a winning definition for the market as a whole on the basis of market share and take up creating a ‘winner takes all’ situation. So the first thing to consider about this announcement is the lack of the major vendors, and that includes even IBM, the big supporter of open source.
Its all to do with the term ‘stack’ where each of the major vendors has tried to produce a complete vertical stack of technologies that are integrated to connect the traditional IT data centric applications and systems at the bottom of the stack with the user centric interactions around services at the top of the stack. This is the fundamental issue to grasp and make decisions around; the top of the stack is focussed on browser / web technologies driven by users to work in different ways in the so called ‘front office’, introducing a decentralised ‘pay on demand’ environment. In contrast the role of the existing IT is to create a centralised enterprise managed environment.
A good definition of ‘front office’ comes from the Business Dictionary; Marketing, sales, and service departments that come in direct contact with the customers, and liaise with the back-office (administrative) departments to maintain a two-way flow of information.
Each of the major technology vendors is trying to make the migration path up the stack the way to go, whereas the new technology vendors, much more frequently XaaS services providers, are taking ‘share’ in the new front office space and offering integration down the stack. So with that thought in mind look again at the OpenStack announcement and the participating companies led by RackSpace, a services vendor, and if it wasn’t for NASA it is all distinctly the piece part players. (Actually I should point out that Intel, AMD and Dell are all signed up as well, but again they are piece part providers in this definition).
That’s not to say the piece parts aren’t good, in fact its arguable with NASA’s Nebula Cloud Computing Platform http://nebula.nasa.gov/ at the centre there is the potential for a real best of breed approach taking shape here. So it’s all about clouds, but by definition that means its all about the web! I might not go so far as the statement by Gary Edwards on his blog where he claims; ‘the battle for the open web is taking place’, but I would like to make the connection a little clearer since it does clarify things in a manner that makes it easier to grasp, and make decisions around. Neither the name ‘OpenStack’ nor the use of open source as an approach is new!
Way back in December 2008 Digg hosted a developer event in San Franciso for the ‘Open Stack’ project, (it was two words back then), which had the goal of making an ‘Open Web’, as the phrase was back then, possible. This translated into a five layer model which you can see neatly laid out in detail on the Yahoo! Developer Network Blog. The five layers offer OpenID, XRDS-Simple, OAuth, PortableContacts, and perhaps the one layer that many people know, OpenSocial.
Open Stack set out to provide a stack for the top down browser / web environment and contains a number of recognisable elements that have already been adopted by a fair number of enterprises. What it doesn’t do is address any of the issues that relate to provisioning services, or hybrid clouds (public / private services provisioning etc). However if you view the topic in this way and grasp what is part of existing web technologies and deployments, and not think in terms of this meaning an entire recreation of a whole stack, then to my mind it does make it easier to understand how to continue the adoption of web-based capabilities and standards as a prerequisite to an open cloud.
The actual wording that OpenStack uses about this is; ‘while many of the components of OpenStack have been in production for years, we are in the very early stages of our efforts to offer these technologies broadly as open source software’. Don’t take this to mean that I am not a supporter of OpenStack, I absolutely am, I just want to help with a practical build on the many excited posts on the topic that portray it as a whole new move. So if you want to go further into the new, and very worthwhile, projects that handle the cloud resourcing elements that OpenStack is lining up then go to their launch pad or you can use their Wiki to get involved.