Defining cloud computing

Dale Vile, originally published on Computing


Published/updated: September 2010

The media, social media, analyst and marketing domains sometimes act like an echo chamber. The same things are heard over and over again on high-profile topics as messages from a small number of original sources bounce around, with detail and precision often being lost along the way.

And so it is with cloud. When you get past the talk about paradigm shifts, market transformation and that cliché, “the cloud changes everything”, you end up with a range of definitions, some woolly, many of which conflict with each other. The only commonality is that it usually boils down to something to do with hosted services.

The net effect is that the more you investigate cloud, the more confused you can become. In terms of specifics, some put the emphasis on infrastructure, and focus on the renting of server and storage space from service providers, often with reference to players such as Amazon. Others focus on hosted applications and speak of software-as-a-service (SaaS), with Sales­force.com typically cited as their darling.

You even come across differing views within individual advocacy groups. Some use the ‘C’ word to describe pretty much all forms of infrastructure hosting, for example, including the traditional contract-based approach, while others say that services must be based on the elastic pay-as-you-go model to qualify as proper cloud.

With application services, the arguments tend to be around the purity of the architecture, and while some say this does not matter, others are fanatical in their belief that multi-tenancy applications are the only true foundation for cloud.

And it doesn’t stop there. Beyond infrastructure and application hosting, we have the platform-as-a-service wars between Microsoft, Google, Salesforce.com and others. This is to do with standards, interoperability and portability in relation to hosted application development and deployment environments. There are then various telco plays around hosted communication services, and a number of vendors and analysts talking about so-called private cloud, community clouds and so on, with reference to more closed shared computing environments.

The word “cloud” has no agreed meaning in the industry, so it is not surprising that many of the people echoing the noise gravitate to the view that “it’s all cloud”, and speak or write in a non-differentiating manner, which can be misleading.

So should we just dismiss it all as hype and spin?

Well, no. The trick is to focus on what is underpinning the noise in terms of fundamental trends and developments. When you do this, it becomes clear that cloud computing in its various forms is not the cause of anything; rather, it is a marketing-friendly word used to label certain effects of the ongoing evolution of capability in some key areas.

Those key areas include developments in the speed and robustness of communications, the coming of age of highly scalable commodity-based datacentre infrastructures, the emergence of flexible and scalable software architectures and, not least, advances in management and provisioning systems. If these are the causes, one of the most obvious effects is allowing service providers to offer hosted capability of various kinds in a manner that is much more cost effective and accessible for the customer compared to even three or four years ago. This, together with new flexible commercial models, is shifting the line in terms of what it might make sense for customers to have running in a service provider environment.

But we are still mostly talking about hosting, so all the things you have had to think about in that context for years still apply in terms of how to manage integration, accountability, support, data security, regulatory compliance and overall service levels when activity crosses the organisational boundary into one or more third-party domains.

Indeed, it could be argued that as hosted services become more accessible, there is a danger of checks and balances being bypassed and organisations ending up with a tangled mess of service contracts, not to mention conflicts, redundancy, gaps and lock-in issues, that may come back to bite them in the future.

Approached responsibly, however, real opportunities are emerging from the confusion. Economies of scale in the service provider domain are bringing down the cost of hosted services in general, making even traditional contract-based offerings that have been around for years potentially more attractive than they once were.

Advanced provisioning, billing and self-service capabilities are providing a level of flexibility not previously seen, creating some brand new use cases around hosting. Consider, for example, overflow scenarios allowing more resources or capacity to be acquired on demand to deal with a transient need, then dropped again with no cost or contract implications. Highly fluctuating web-based applications, periodic workloads such as payroll and billing, and IT requirements around development and test environments spring to mind when we think about the relevance of a more elastic approach to hosting.

So there are tangible benefits from looking more closely at the specifics of this thing called cloud. We just need to remember that it’s not some kind of magic – it’s just IT.




Featured Content