Every which way but loosely coupled

One of the most appealing aspects of self-service 'retail' style clouds such as Amazon EC2 and Flexiscale, is that they simplify the process of deploying 'software as a service' applications and web applications generally. The Amazon guys frequently cite Animoto as a case study - without any capital investment and without paying consultants, the tiny Seattle-based Animoto team were able to launch, and grow to a substantial paying user base, in a very short time.

By the opposite token, IBM have quite recently announced more cloud initiatives. More precisely they have announced some data center outsourcing plans and consulting services offerings, and labelled them 'cloud'. Take a look.

This is very carefully worded press release. Look at how many times 'cloud' is mentioned in this press release without saying that there actually is a cloud. Here are some examples:

1. "cloud computing capabilities" - ie. it is capable of being a cloud but may not be one
   
2. "cloud-like computing mode" - cloud-like but not a cloud
3. "capable of supporting cloud environments" - but is not actually a cloud
   
4. "cloud computing center in Tokyo, Japan will provide large enterprise customers, universities and government agencies immediate access to experts who can help them deploy cloud computing environments" - buy expertise, i.e. time and materials, then buy more people who install some software
   
5. "Cloud computing gives organizations the opportunity to remotely access a vast network of computers" - a general statement
   
6. "cloud center will be linked to the new Raleigh center and IBM's seven other cloud centers throughout the world, to help clients pilot cloud infrastructures" - pilot but don't use
7. "IBM offers a number of solutions to help clients deliver cloud-enabled services to their customers -- an initiative dubbed Blue Cloud" - not sure who is selling what to who here but it sounds like consultants are involved
   

And so on and so forth. Now I realise that IBM have good reason to embrace 'cloud' by extending their data center outsourcing and global services businesses. And, I fully appreciate that IBM might like their enterprise customers to feel that piloting cloud-like deployments in safe IBM data centers, some time in 2009, is the way to reduce risk. After all, IBM are all about reducing risk, right?

I also applaud IBM's stance on development of geographies possibly not yet so well addressed by cloud, hosting, data center management, and outsourcing services. And similarly I am impressed with statements about energy reduction and cost efficiency, and I look forward to one day seeing their green API. Because the more green tape there is, the more attractive it will be to use a certified carbon neutral data center, rather than build your own.

But what I don't understand is why they keep talking about the future, when self-service clouds exist today that do not require "a set of hardware, software and services that allows IBM clients to offer personal and business services from remote, centralized servers". And more are appearing.

Yes, by all means create a market for 'solutions' around cloud vs non-cloud vs private cloud deployments. Yes, go ahead and sell professional services for migrating applications into service-oriented, distributed, and virtualized, infrastructure.

But blue as it may be, in what way is this 'cloud'? It's not 'on demand' if I cannot do it myself. Surely the whole point of the cloud is disintermediation - the Animoto on EC2 use case did not require the interposition of experts, risk managers, hardware sales, or software installation. Large enterprise customers will of course have more complicated needs than Animoto, but enterprises still want the fundamental benefits of cloud - pay for what you need, and reduce resource when you do not need it. Less is more. Loosen up.