Currently my life is full of puzzled people from various sections of the IT community asking for enlightenment on ‘cloud technology’. Their puzzlement usually stems from the simple fact that they can’t make a whole lot of sense about when, and where, they would gain by deploying cloud technology in their current operations. A barrage of questions then follows, mostly about their doubts over security and reliability if they were to take important legacy systems outside their enterprise into a shared environment. Then comes the inevitable question, does this mean they should be building an internal, or private, cloud? And so it goes on….
My usual reply is quite simple. If you evaluate cloud technology solely on the business case behind replacing the existing technology operations layer, it’s hard to see the value. That’s not to say that you can’t make a great case for ‘virtualisation’ to improve internal utilisation of your current investment, as you certainly can. But that’s a different argument to the one most people are trying and failing to make around their perception of clouds as a new external enabling environment.
So why be interested in clouds at all then? Again there’s a simple answer. It is to provide for and support a new generation of business requirements which are built around ‘services’, as in SOA and not around applications of the type IT is responsible for running today. These new requirements are already becoming visible, but not always too clearly as they are often not taking place in the IT department at all. Many of them are already provided by popular hosting providers such as Google, Amazon, or Force.com. However, what is clear, based on a number of articles and blogs I’ve read recently, is that users and business managers alike are becoming technically aware enough to ‘assemble’ software in such a way as to deliver what they want.
The term ‘shadow IT’ is something that my colleagues and I identified some years ago in our book MashUp Corporations, and was more recently identified by Gartner as ‘citizen developers’, in a great presentation by their analyst Eric Knipp. Eric’s claim, which is acknowledged as controversial and not a formal Gartner opinion, is that 25% of new business developments by 2014 will be of this type. He offers the following definition:
“A true ‘end-user’ developed application is typically an ‘active document’ created as a natural process in an employee’s day-to-day responsibilities. Multi-user line-of-business applications are based on distributed computing principles and are typically created by IT. LOB application development is a technical pursuit outside the grasp of the skills normally found in business units.”
I would press for this definition to be extended to encapsulate ‘situational apps’ the topic of an earlier post this year. I’d then go on to argue that this kind of ‘citizen-led’ activity can only really flourish if there is a total abstraction from anything to do with the technological consequences that underlay the situational apps. Now this is the crux of the debate about clouds, what they can deliver and why they will introduce a real transformational change in enterprise working practice. It’s not a question of ‘private or public’ clouds, or similar location terminology about where and how the service is delivered. It’s not even about how it is paid for. No, it’s all about the ability to allow business users to assemble and use the business ‘services’ that they want, when they want them, without having to think about the problems and issues that this might cause at the technological level.
Extend that thinking and it becomes obvious that the edge of the enterprise at the firewall is yet another technology-defined barrier and that ‘services’ from outside could also be required. How many services, by how many people, and how often? That’s an IT application question about provisioning back office applications paid for as an annually budgeted overhead and using them in a very controlled and defined manner. In the new business technology model enabled by clouds there are no such limitations, instead users in occupations built around doing business outside the enterprise – sales, purchasing and supply chain for example – will consume what they need, when they need it without technology constraints based on payment against ‘consumption’. We are talking about a payment model that gives direct cost attribution to the user or business unit concerned, in short the SaaS model (a model that is also the subject of a lot of hype and misunderstanding about when and how it is applied!).
A key part of this debate is the difference between information technology enabled by the technology revolution of the late eighties and early 90s, and business technology, meaning something different and enabled by the recent technology revolution of the Web and clouds. IT is focussed on supporting internal, back-office, core transactional processes that automate core enterprise activities and have clearly defined characteristics in terms of numbers of users and volume of activities. Moreover the funding model is by means of annual overhead budgets and recovery. Business technology by contrast is an externally-oriented, front-office capability with enormous variations in use, both as to users and volumes, and needs to be paid for by those who use it when they use it.
The role of cloud computing is to address this requirement by its ability to provide a completely different provisioning model, both in terms of how it uses technology – separating the users from the technology issues – and by its funding model. Virtualisation is a major part of the technology aspect for sure, but it’s not used in the same way as applying virtualisation to selected resources in IT in order to reduce cost and improve utilisation of those resources.
So there it is – a controversial point of view to try to make a simple point. Owing to the hype factor around cloud computing, one would be forgiven for thinking that everything everywhere should be ‘in the cloud’. I’d like to propose a sea change – let’s end the hype with some good common sense so we can focus on ensuring a new technology is deployed for the right reasons and produces really good results!