Understand Cloud Computing in basic term

Cloud computing

you are here because you want to understand cloud computing. When I say cloud computing there are some basic questions which will come into mind these are :

Why cloud computing ? 
Why are people talking about the cloud ? 
why is it an important investment ?
why does it seem like such a big shift in terms of our notions about computing and deployment of our applications ?

So lets start here and try to explain in simple term

Traditionally we've had the on-premise model where we buy all of our hardware up front, we pay for full-time employees to manage that hardware and the software installed on it, and we take all those costs up front and annually to maintain those things and to pay those salaries.

Then organization started looking for more flexibility because they wanted to offload some of these things in hosted arena. They wanted to remove the overhead of power consumption, bandwidth, management,full time employees, hardware and support that goes along with that, so they started moving to leased model.

Now this doesn't necessarily mean that we don't also have costs for resources, full-time employees, to manage the applications and things that are installed on there. We still have people that have to manage the software packages we install on there, maybe they're also responsible for applying patches to the operating system.

Here comes cloud, 

And finally with the cloud, we move away even from a fixed amount lease to this notion of metered access where we have this unlimited number of resources, we don't necessarily have physical servers that we're tied to, we just have virtual instances, and we're going to pay as we go.

Consider below situation :

I need five servers right now to run any application ,we are going to use those for a month and then we may take those back offline and stop paying for them.

No need to worry, we have a lot more flexibility with cloud computing, as our needs change, as our application scale changes, as our usage changes, as our requirements change. The work of managing that infrastructure and managing that hardware is pushed off to the cloud provider.

It helps us eliminate risk. how ? 

We can think about what happens if we over-provision. we get too much capacity, we buy too much hardware,too much bandwidth. it means we're spending money that we don't have to spend because we don't have the demand to keep up with it.

If we under-provision, if our capacity is below our demand over time,it may be possible that we have periodic downtime where our system is not able to handle the load and customers get frustrated because our app doesn't respond.we'll likely end up losing customers and losing money. It means our demand is not meeting our capacity, but that's not really what we want.

What we want is our capacity to be able to scale up with the demand. So we want to pay for the use rather than having to provision for the peak amount. Instead of having this static data center with a fixed number of servers, we should be able to watch the demand, as the demand increases very quickly capacity should be flexibly increases. As we start getting more users and we start needing more resources, we want to simply go in to a portal and provision them.

I don't want to go and fill out a purchase order and buy a new server, get it all installed and patched and ready, that's going to take time, it's going to take resources. I want the concept of the cloud and the scale where I can simply go in and provision a new node, take advantage of that elastic infinite resource out there to meet my demand. 

When you do this, and when you use those data centers that are out there, you get the economies of scale, you get benefits from the network bandwidth, the storage, the administration, all of those things end up costing less when we do them in bulk like this.
When we have huge data centers that Microsoft owns or Google or Amazon, the individual cost of those things goes down, which saves us all money.

So as we think about the cloud platform, we really have these three key things that are new.

1. The infinite computing resources that are available on demand where I can go in and simply say I need another computing resource node. Maybe I have 5 right now, my demand is up, I need 6 or I need 10.

2. I don't have to pay out money up front for those resources. I use only what I need, I pay for what I need.

3. We have short-term billing. The billing is generally on a monthly basis. So if I have a month where I have high usage, I'll pay more for just that month. If my usage and my capacity demands go down the next month, then the cost will go down as well in those shorter cycles, so I'm not leasing or paying for a year or five years, I'm paying on a much shorter cycle, and again reflecting the actual usage that I've had of the resources.