What is Cloud Computing?

By Alan Lindsay
Updated: May 25, 2012

What is Cloud Computing?

Cloud computing may be one of the hottest topics  for businesses of any size and even for consumers but it is hard to find out what people mean when they use the term and in addition different groups really do mean different things when they say it. So just what is cloud computing?

Here are some definitions:

'Enterprise computing services packaged to meet the needs of medium and small businesses.' (Forrester)
'The delivery of computing as a service rather than as a product.' (Wikipedia) [http://en.wikipedia.org/wiki/Cloud_computing]
'The dynamic provisioning of IT capabilities (hardware, software, or services) from third parties over a network.' (Accenture) [http://www.cio.com/article/501814/Cloud_Computing_Definitions_and_Solutions]

While there is commonality between these various definitions, they are not the same and they are still incomplete. This is because the many vendors in the 'cloud computing' space actually also offer overlapping but different services and solutions. You variously hear things like software as a service (SaaS), infrastructure as a service (IaaS), virtualization, or even simple internet services all described as cloud computing. Each of these, and each of the other services described as cloud computing, is a piece of the answer but not all of it. You even see consumer services like Gmail or Google Apps or Microsoft Skydrive or Dropbox get called cloud computing - and in a small sense, they are.

But before we settle on what cloud computing is, let's take a look at why it is important and how it came about. Cloud computing is important because it offers businesses of any size a way to reduce the complexity and cost of any and all aspects of their computer systems at any level from hardware and networks to software and services. Computer systems are one the most complex and costly aspects of any modern business. And that is before taking into account the people costs of managing and maintaining those systems. Outsourcing the equipment, software, services and personnel is a great way to reduce that complexity and cut costs.

But how can cloud computing vendors manage to do that? If you take your company's servers and stick them somewhere else they are still the same servers and still cost as much. Same with software and same with the people and time to support them. That is where virtualization comes into play. Rather than your organizations four servers (say) and another organizations four servers and so on, the cloud computing vendor supplying infrastructure as a service runs something like six servers for all three companies and virtualizes them into being a pool that can act as all twelve servers if needed. This technique relies on the fact that not all the power of all those servers is needed at the same time. So the vendor can save money on hardware and pass that saving back to its customers. Same principle for software and for services and even for support staff.

This gets us closer to an understanding of what cloud computing is. To the cloud computing consumer (whether an individual or an enterprise) it is the provision of computing power and software and services remotely and in a transparent manner. That means that the compute power and services have to scale up and down to match demand. If the company in question wants double the power it should be made available. This is where many of the practical details of the differences between cloud vendors surface - in pricing models and service level agreements and in the amount and timeframe in which it can scale to meet demand. But to the vendor or to the consumer who wants a much more granular approach, cloud computing also involves the physical location, the precise servers and services offered, the level of virtualization and the amount of control over these that is exposed to the control of the customer.

And, to complicate things further, for each of these kinds of cloud service there are several levels at which they are provided as a service and at which they are virtualized. Take something as straightforward as web hosting. If a company wants a simple web presence they can get a shared hosting service very cheaply where they are simply buying a defined small share of a virtualized web host running on a virtualized server at some nebulous and likely undefined location. Or they could get a specific and dedicated server and host instance that is not shared in any way and that might or might not be virtualized. At the other end of the spectrum they could simply have the provider host their own servers at their location - essentially paying for connections and power and maintenance. Or almost any imaginable level of scale in between those two scenarios.

It is easier, as the Forrester team tried to do, to turn the question around a bit. What does cloud computing do for its intended consumers? It abstracts and virtualizes the need for as much physical computing as possible and then it dynamically provisions that computing so that it can change to meet user demands. Finally it is provided at some level of service guarantee with defined levels of capability and cost. In other words, it is a leasing model to provide an organization with computing power rather than a purchase model. The cloud in the name is a nebulous fuzzy 'cloud' of computing hardware, software and services that are used to meet the end user need.

Looking at it this way it is clear that many cloud computing vendors spend a great deal of time talking about what their capabilities and service and needs are - in other words HOW they provide cloud computing - rather than the precise services they offer their own customers - WHAT they offer. Since we believe that the important part of the equation is what they offer, this is how we define cloud computing. Cloud computing is the suite of virtualized computing power, software and services offered to the end customer. At the consumer level, say Hotmail, this is a free and unlimited email service with guaranteed backup, access and storage. At a slightly higher level, say Google Apps, this is unlimited access via the internet to a suite of basic office applications including storage and tools to make them useful (for example multiple user access and editing to a single document). At the other end, for enterprises, this is the complete provision of computing, applications, services, storage and web hosting, relational database, commerce services, and more.

Related Categories
Featured Research
  • The Basics of Real Time Personalization

    Real-time personalization of the customer experience has been described as the holy grail of digital marketing. And the race is on. Gartner believes that by 2018, businesses that excel in personalization will outsell those that don’t by 20%. Though the benefits are clear, the path to get there is not. more

  • The Benefits of Cloud Computing

    Cloud computing represents the next phase in the logical evolution in the delivery of IT services. This complimentary white paper from IBM will explain how cloud computing will help your business through specific examples from cloud implementation abroad, and more. more

  • Defining a Framework for Cloud Adoption

    This complimentary white paper from IBM introduces the cloud computing adoption framework from IBM and make it available for use by any organization looking for a standardized frame of reference for cloud computing discussions. more

  • Adding Networks to Integrated Service Management

    Integrated Services Management, or ISM, is a major evolutionary force guiding management technology and practice progress in enterprises today. The implications of ISM are broad, but the essence of the initiative is a re-thinking of how IT delivers value to the organization it serves, and a re-framing of all IT contributions around the concept of services. more

  • Gartner Magic Quadrant for Operations Support Systems

    With IBM's Service Assurance and Fulfillment solution CSPs are able to integrate business processes with service delivery systems, modernize legacy and proprietary systems, and focus on fast return on investment. And in the first ever Magic Quadrant for Operations Support Systems (OSS), Gartner has declared IBM the "leader" in the space. more