I had thought that in the month of December that there might be a slow down for the Technology Industry, but seems not so, or at least if the topic is Virtualisation or Cloud Computing. I did get to go to the HP Software Universe event in Vienna, but I didn’t get to go to the grandiose titled ‘World Summit of Cloud Computing’ held in Israel, but I did follow the proceedings remotely with interest.
Once again it restarts the challenge of what is Cloud Computing versus Grid Computing, and as the sponsoring organisation for the event is www.grid.org.il you can guess that a lot is dedicated to the Grid side. However there is a really good, and informative, kicking off point on their web site that provides detailed definitions and lots more on the grid side and though it appears to offer a similar service for the cloud computing side it sadly does appear to transfer it back into Grid Computing again. I quote:
Cloud Computing – Pay-per-Use for On-Demand Scalability
Over the last few years we have seen grid computing evolve from a niche technology associated with scientific and technical computing, into a business-innovating technology that is driving increased commercial adoption. Grid deployments accelerate application performance, improve productivity and collaboration, and optimize the resiliency of the IT infrastructure.
Why does this bother me? After all if it works to reduce costs in infrastructure provisioning isn’t worrying about the terminology a semantic game for those with a personal view to express? Well, fortunately for my concern the following lines appear as well;
Today, the maturity of the Virtualization technologies, both at the VM and at the IT infrastructure levels, and the convergence of the Grid, Virtualization and SOA concepts, enables the business implementation of the Cloud Computing for utility and SaaS services.
You will notice that I have put SOA in highlights, and here is my point, that there is a big difference to understand between the Cloud Computing/Grid layer and the layer above providing Cloud Services. If you want to take the current applications, data, etc and buy some operating capacity using the same application format then its Grid Computing for sure. However when you consume services such as Google maps then you are using Cloud Services, and the repetitive use of word ‘services’ is no accident! Let me try to explain something to get the point across about how different this world is; you don’t get a file as in the conventional sense when you get the map, you get a ‘representation’ of the map on which you can position other representations of services. Cloud Services are stateless and loose coupled services completely different in every way from conventional applications and their integration which are in comparison both statefull and close coupled.
So if you are expecting to interact externally with customers and suppliers across the Cloud, (as opposed to transact internally with your existing traditional applications), then you will need to present your services to others, as services, and be able to consume their services in return. All of which can sit over the flexible resource model of the cloud computing layer underneath, as once again unlike a traditional application there is limited control over the numbers of users and the pattern of usage. All of which brings me to my key point; you will need to have adopted SOA internally and properly, to be able to connect up externally to the Cloud Services. This positions the role and importance of SOA in a whole new light!
As the action focuses on the Cloud Computing layer with MIPS around utility pricing it certainly grabs the interest, especially under the current economic conditions, but the lack of information about the overall development of Clouds is really worrying. So much so that Russ Daniels, the CTO of Cloud Computing at HP, and myself teamed up in the summer to try to write an informative paper on the topic, with no mention of products, just an attempt to tell the whole story, with a roadmap from implementing SOA today, through to the ‘services’ model of tomorrow. And most importantly how the definition of those services will be based primarily on business tasks encapsulated into services rather than in many SOA implementations were the focus has been on technology tasks.
I will be presenting from this paper at HP Software Universe at 10.15 on the 11th December in room 01/J241/1 under the title of ‘Cloud Computing; hype, business enabler, or infrastructure cost play’. Funnily enough Russ is at the Cloud Computing Summit so can’t be there. The Cloud is, of course, all three elements, and I will try to dismiss the hype, to show how to get the other two with a roadmap to get you there. Come along and join in the session, I would love to see you if you are there. If you can’t make it then you can download the paper.
Lastly, a real gem of a find that might help us all! Take a look at James Urquhart’s excellent blog posting on a ‘bill of rights for cloud computing’. Great stuff thanks James !