The whole world has heard about cloud computing – even if there does seem to some confusion about terminology. According to a survey from Six Degrees, 83 percent of customers think that cloud service providers could do more to demystify the cloud. And that’s not to mention the survey that revealed American users were even more confused and thought it had something to do with the weather.
Why is there this confusion? The US National Institute for Standards and Technology defined the term about three years ago. "Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a sharedpool of configurable computing resources (eg networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.”
What is cloud computing?
To put it another way, the fundamental idea behind the concept is to rent the provision of a service from another company, and for that service provision to be delivered via the Internet: the essential core idea is that you do not run central services inside your organisation, but push them out to these third party suppliers.
In some ways, it’s similar to the long-established concept of outsourcing but there are differences. There’s no fixed contract length, services are on a ‘pay-as-you-go’ basis and there’s a strong element of self-provisioning, all changes that mean many decisions are taken out of the hands of IT departments and delivered to the users themselves.
It’s not a concept that’s entirely unknown to us. As consumers we’re used to all sorts of cloud services; email like Gmail or Hotmail, video sites like YouTube, storage services like Dropbox or iCloud, and social media companies like Facebook or Twitter. None of the software for any of these companies resides on the consumer’s own devices but we’re all happy about using these services.
And services is the key word; a fundamental aspect of cloud computing is the move from talking about hardware and software to talking about services – it’s a big shift.
Cloud is usually defined in one of three ways: infrastructure as a service (IaaS), platform as a service (PaaS) or software as a service (SaaS). IaaS refers to the hosting of a cloud service, PaaS is where cloud developers play and SaaS is cloud-based delivery of software. There's a bit of a blurring of lines on these - particularly between IaaS and PaaS - and some commentators have wondered how long these distinctions will hold but they serve as useful shorthand.
Cloud services
What services are provided? Well, at the simplest level, you might be renting server space, usually contained within a virtual machine environment. More likely though, you will be renting use of a service – an email server, a web engine, or an application hosting engine in which you can run your own custom applications.
Why would you want to do this? Well, there are several reasons. Many of the initial early adopters chose cloud because it was seen as an easy way to save money. While it’s true that in most cases cloud will cut costs, it’s not necessarily a given. However, the driver for cloud these days is more often the need for flexibility and the requirement to provision services quickly.
There are other factors too. One of them is the economy of scale. A datacentre can provide access to computing power at a lower cost than you can provide it in-house. There is one large air-conditioning system, one big power supply, one set of security and so forth. It’s not just a question of economies of scale, it goes further than that: if you are renting a certain number of seats worth of, for example, Exchange Server, then you don’t need to worry about making sure you have bought the right size server. Choose something too small, and it will struggle. Too big, and you have wasted money. In a per-seat rental environment, it’s not up to you to worry about the sizing. The costs are entirely predictable, with no surprises.
This all sounds good but there are some even bigger advantages. You usually can forget about having to worry about keeping the server patched and up to date. That’s down to the service provider to do, and this can take away a considerable headache and cost. You will also benefit from the scalability of cloud too – if you current internal server on which your email server dies, then you have to get a replacement up and running as soon as possible. But this could take days.
In a cloud environment, your email server is running on one of maybe thousands of servers. If “your” hardware fails, your workload can be quickly moved to another box, and this happens automatically. This is not a trivial concern, the typical IT department spends 70 percent of its time keeping systems up and running – by being freed up from some of the routine maintenance, there is more of a chance to use IT creatively and add real value to your business.
Finally, if you are running a system which has an unpredictable load pattern, for example an e-shop in the run up to Christmas, then you can have real problems scaling up the capabilities of a local server to cope with the unexpected load. Worse still, the internet connection into your office might get swamped. With cloud, not only will the datacentre have a huge amount of connectivity, but it might well be possible to add in more computing power for a few days to cope with the load. This is called scale-out, where you bolt on more horsepower for the days when you need it.
Of course, going with cloud is not a universal panacea, and like any IT implementation, going with cloud requires careful planning, for example, it’s essential to make a careful assessment of how it will change the load on your existing infrastructure. But a well-designed and implemented cloud deployment can be a radical improvement to a company’s IT infrastructure, which is why there are so many companies making the leap to cloud.
Getting the best value for money out of the computing infrastructure has always been a target, especially now when budgets are tight – cloud computing offers one way forward.
This is an updated version of an article orginally written by Jon Honeyball and published on 28 March 2011
