Most of you are probably passingly familiar with "The Cloud" being something at the moment. It tends to crop up in advertisements, IT articles, mobile phone marketing... anywhere that regularly reports on the IT sector will have it crop up a few times a week. You should probably bring an umbrella.

What Is The Cloud?

The Cloud is, in its simplest form, an amorphous blob of raw computing power that can be delivered to customers in many forms as a service. The Cloud itself can be made up of tens, hundreds or thousands of high-powered computers known as "servers" which are housed in a "Datacenter".

We will look at these first.

Servers and Datacentres

A "server" is essentially any computer that has a shared resource over a network or the Internet. A "dedicated server" is in use all the time for that purpose. The servers we will look at here are the second variety. The ones in use in a Cloud environment are designed specifically to produce massive amounts of computing power, use as little space as possible, and run 24/7 without any manual intervention. One such design is known as a "pizza-box" or "rackmount", since it is a thin flat square. These are designed to be stacked vertically in a rack to collectively produce silly amounts of raw power.

The servers that generate this computing power are kept in a specially-designed building called a Datacentre. These are climate-controlled, dust-free environments with uninterruptable power sources, sophisticated ventilation systems and are designed for the express purpose of keeping a healthy environment for server to continuously operate in.  They also frequently look cool.

All of this is designed for one purpose alone: to guarantee the availability of service 24/7 to those who are paying to receive it.

Notable mention goes to Google, who apparently keep their servers in the dark so that naughty server technicians can't see or steal any of their custom-built hardware designs and sell them for giant amounts of money.  They also use toilet water as coolant in the server cooling system. Clever?  I guess so.

Computing As a Service

To begin to understand how the cloud operates, let's coin an analogy.

Think of cloud-based computing power as being similar to electricity.  Power companies generate electricity using not one, but many separate generators.  These generators collectively supply the electricity and are housed inside a specially-designed building: a power station. This electricity, as a commodity, is delivered over the national grid directly to you without you having to own any of your own generation equipment.  So long as you are connected to the grid, you can contact a service provider, tell them you want electricity, and voila... you receive it.  You can use it however you want... power hair-dryers, televisions, microwaves and so on. They send you a bill depending on what package you are on... you may pay a fixed fee, or it may vary depending on usage.

Now replace a power station with a Datacentre; power generators with high-powered servers; the electricity grid with the internet; electricity with computing power. This is Cloud-based computing in a nut shell, the delivery of computing power as a guaranteed 24/7 service.

What Does The Cloud Do?

The analogy, while being a fairly close approximation of what is going on, changes slightly when it comes to the receipt of the service. It goes without saying that electricity and computing power are not the same... you are not powering your television with computing power; it isn't coming out of a tap in your house or heating radiators.

The specific computing power or product you receive can come in different forms depending on what you are buying.  Here are some examples:

Online Storage

You can store anything you want safe in the knowledge that it is guaranteed to be there at any time you want to retrieve it.  You might have 1GB of space, 100GB of space, or unlimited space.  It all comes down to what product you took.  You can then access your files anywhere that has an internet connection, and so long as they are in the Cloud, there is no chance they will be lost.

See here (JustCloud) and here (Google Cloud Storage) for real examples of Cloud-based storage services.

Phone Content Storage

Get a new phone, insert your SIM card, input your username and password for your Cloud-based account, and it synchronises all your contacts and content instantly.  Then replicate that same content onto another 10 phones by linking them to that same account.  Any time you add a contact, it is created and stored on the Cloud account.  Lose your phone?  Damage it?  Don't worry... all your pictures, videos and memories are safely backed up to the indestructible Cloud.

See here (Apple iCloud) and here (Google Sync) for real examples of Cloud-based phone storage services.

Virtual Private Server

This is the most potent and complicated product on the Cloud.  From the amorphous blob emerges your very own perfectly formed standalone virtual computer (server).  You specify how much processing power, system memory and storage capacity it should have, and the provider supplies a virtual server to that specification.  Of course, you cannot sit at this computer... it is not a physical object.  It is created purely as an abstraction from the huge raw pool of power in the Cloud using specially designed virtualisation software, but it is designed to act to you and the world exactly as a physical server would.  You administer it remotely from your own computer using an interface that allows you to control it as if it were infront of you. You can realistically do anything you want on one of these... run a single or multiple websites, run an online shop, a gaming server, use it for personal or business use... the list is long.  The server will always be there, there is no hardware to fail, and no maintenance to be carried out. Like any service, you pay for it, and it is always there.

See here (Fasthosts) for a real example of a Virtual Private Server provider. Inverclyde Now runs on a Virtual Private Server with 4 processors, 4GB system memory and 80GB of disk space.  While this is a fairly high powered machine, it is a mere atom relative to the size of the larger Cloud.

Overall, with cloud-based computing, you are using a tiny slice of the Cloud for yourself... and doing more or less whatever you want with it.

Why? What Are The Benefits?

I will demonstrate the reasons for adopting the Cloud with a real world example.

My present job is as an IT Systems Analyst for a company based in Glasgow.  We have offices all over the country, with over 1,000 staff and 530 computers. Our IT infrastructure uses a model that is known as a Thin-client environment.  I will quickly explain what this means before going further:

Thin-client Computing involves the use of what can be considered a "mini-Cloud" for private use by my company's staff.  We have five extremely high-powered unmanned physical servers in a single location, all configured to function as a single collective resource.  We own these and maintain them ourselves inside a Datacenter.

Users around the country have "thin-client devices" to do their work which are little more than small black boxes that contain a single circuit-board.  They have no computing power of their own.  When you switch them on, they connect to our mini-Cloud via the Internet and act as a terminal to the mini-Cloud, where the actual computing and processing is carried out.  They are useless without the mini-Cloud to do the hard work.

From the user's perspective, even though they are accessing a system which may be hundreds of miles away, they see it on their screen as if it was a computer in the same room.  The devices are called "thin-client" simply because they are lean, thin and inexpensive.  The mini-Cloud servers are refered to as "fat-clients" in this environment, they are expensive, powerful and can support the work of hundreds of remote users at the same time.

Comparing the cost of five massive servers and 1000 thin clients to 1000 normal computers... it becomes apparent why this model is prefered.  Thin-clients are cheap to buy, use hardly any electricity and, through the mini-Cloud, can do anything ordinary computers can.

Now that we've covered that... let's look at the mini-Cloud.  If this goes down then nobody can get any work done.  It is therefore business-critical that it remain online 24/7.  People can work from any location with Internet access at any time they want by connecting to the mini-Cloud.  It must always be there.

Until now, my company has maintained our own hardware to run the mini-Cloud.  We purchased and maintain the five physical computers. This involves a huge spend on the hardware, which has a shelf-life of around five years after which the spend must be repeated.  During those five years, if any device or piece of hardware breaks, we need to repair or replace it.  We need to keep our own data backed up.  We are responsible for continuity of this system that the organisation depends on every single day.

Returning to the electricity analogy, we are effectively generating our own power.  If the generator fails, we need to replace it.

But... why not just buy it from a company that specialises in supplying this as a guaranteed service and pay a nice fixed monthly cost without us having to worry about maintenance?

Now there is no need for unexpected repairs or replacements.  The Cloud platform is the responsibility of the provider, like the generators to the electricity company.  It is financially easier to plan for this way.

Our new virtual (not physical) servers will run on a platform that is backed up, backed up again, and backed up on top of that without us having to do anything.  If one server in the huge cluster fails, another 10 are waiting to take it over.  We will never even know it happened.  If the Datacenter is hit by a meteor, there is a whole other Datacenter in another area of the country that can recreate our virtual server from data backups within 30 minutes.  Disaster recovery becomes easier this way.

If we ever hit the capacity of our five virtual servers... because they are abstracted from the bigger Cloud, the software that creates them can increase their resource to effectively grow another processor, or add more system memory. Future-proofing is now easier as well.

Anything hosted on the Cloud is reliable, cost-effective and has no single point of failure.  So... that's why.

Finally... Why Is It Called 'The Cloud'?

Network diagrams over the years have historically displayed any collective or complex infrastructure, one that appears to the user as a single resource, using a nice picture of a cloud.  The Internet was the most common example over the years; even though it's a huge and complex infrastructure, to the user it is simply "the Internet", and so it appears as a metaphorical cloud with that name.

Here is an example of a basic diagram that would be a nightmare if the inner-workings of the Internet were explained in detail instead of just using a Cloud:

See how nice and easy that is to understand?  It's the Cloud!