Pinning the term cloud computing down to a single definition is not just hard, it's impossible, primarily because vendors have habitually used and abused it to refer to so many different things.

Over the past couple of years we've have tried to define some of the underlying principles that will remain once the marketing people get bored of the term and move on.

We have developed a way of walking people through the confusion, both interactively and through advisory material, in order to help them figure out how cloud computing in its various forms might fit into their own plans and activities.

More recently, we have turned our attention to looking beyond cloud at the long term strategy and direction level, so we can identify exactly where the opportunities are for CIOs looking for short to medium term advantage. In order to do this, it helps to distinguish between the different types of cloud proposition out there. We have identified three types of cloud proposition that we have concluded represent the best bets for the shorter term:

Resource pooling technologies
The IT vendor community has been working for a decade or so on more capable provisioning and management systems that allow IT resources to the utilised in a more flexible and efficient manner. The basic idea is that you make lots of physical assets, whether they be servers or storage devices, look like a single virtual resource pool capable of supporting a variety of virtualised workloads available on demand in a flexible manner.

Applications and workloads can be provisioned rapidly, overall asset utilisation can be increased, and management overheads can potentially be lowered. Over the past couple of years, the term private cloud has been coined to refer to technology solutions in this space.

Whatever you call it, what we are seeing today is the culmination of many years of research, development and early adopter testing. While some of the jargon is new, the technology being offered in this space is now mainstream ready. It's robust, usable and available at a reasonable cost.

Infrastructure hosting
The cloud speak here is Infrastructure as a Service (IaaS), and the tendency is to focus on the dynamic and elastic services brought to the fore originally by Amazon.

However, traditional fixed contract hosting is still very relevant, so there are two important developments in this space that stem from service providers having access to similar technology that underpins private cloud.

 - Better technology has allowed providers to manage their own resources more efficiently. This is shifting the lines in terms of economics, bringing down prices and making even traditional fixed hosting more attractive in a wider range of scenarios.
 - For customers that need it, elastic services with physical commercial flexibility are becoming more widely accessible on a fully supported basis.

The bottom line is that the infrastructure hosting market, while several decades old, is now throwing up some very interesting new options.

Utility style SaaS
The tendency in the industry is to lump all forms of hosted application services into the same Software as a Service (SaaS) bucket. It helps to distinguish between services that are generally implemented by customers off the shelf and services based on application functionality that needs a lot of configuration, tailoring and integration work to get up and running.

All indications are that the market for the former — straightforward solutions such as email, basic content management, social business, unified comms and collaboration represents a stronger short to medium term opportunity than the latter.

The utility nature of these propositions translates to rapid access to new or improved capability as well as potential cost savings. We also see lots of investment within the consulting community to help organisations with their utility-style SaaS initiatives.

Following on from this last category, more complex SaaS offerings such as ERP and broader scope CRM, beyond simple sales force automation, involve a similar level of evaluation, due diligence and implementation work as their traditional on-premise equivalent, and the potential for cost saving is generally lower.

Furthermore, unravelling the issues around information security and management, development, maintenance and support of interfaces that cross the devide between on-premise and provider service can be quite a challenge. Separating utility and complex SaaS offerings avoids organisations rejecting the former because of issues that really have more to do with the latter.

The cloud proposition we haven't mentioned yet is Platform as a Service (PaaS), which is basically about hosted application platforms upon which custom cloud applications can be built and executed. It is beyond the scope of this article to discuss this fully, but suffice it to say that we see this as the least well defined, and therefore the most risky, of the cloud plays.

While we would encourage organisations to gain experience in this area sooner rather than later, and potentially take advantage of PaaS for tactical requirements, making strategic commitments is hard when standards are immature and competition between proprietary services has yet to play out.

So, for the time being, we would provide three recommendations to CIOs in relation to cloud:

 - Pay attention to it as there is some interesting stuff going on
 - Avoid binary thinking; it is less about evaluating cloud as a whole, and more about looking at the opportunities represented by the discrete categories of offering
 - Focus on the three propositions we have called out when looking at short to medium term adoption

The trick with cloud is always to be clear on the specifics.

Pic: Kevin Dooley cc2.0