Where the cloud make sense @aws

Hi again!

Today I want to talk about the cloud and when cloud make sense. Today no just words, let’s see with a real case.

As you provably know, the cloud really exists. There are companies like Amazon AWS, Rackspace and others who offer their service of hardware renting for your personal use. What’s the difference between this cloud companies and your normal hosting company who can rent their servers? At the first time you can think that there is no difference. When you take a look more closely, you start seeing some differences, some details and added services that a normal company doesn’t offer. Enough important to take the decision to move to the cloud and not to a conventional hosting service.

The most important differences between each are the following: Data migration, data replication, fault tolerance infrastructure, fast recovery services and other value added services that hosting companies doesn’t offer like CDN, distributed file systems, etc.

As you can see there are a lot of services that a small/normal hosting company couldn’t offer. For this reason cloud services are a really good option to deploy your infrastructure.

Let’s see the real case where the cloud make really sense. In this graphic you can see an abrupt increase of demand. This case is a real case of a production platform.

Captura de pantalla 2013-03-14 a las 14.47.31


In a normal histogram the line rarely grow over 15. What happens if you receive this sudden increase of demand? In a normal infrastructure you provably offer a bad service or directly not offering to your clients. In a cloud infrastructure you just deploy more instances and make this work done it.

This case is based on a encoding video queue. Unexpectedly you can receive an big amount of demand, in a normal infrastructure, you can’t do anything. In a cloud infrastructure, you can try to deploy a new instance and balance demand over the new instance. This is really interesting when you could have unexpected spikes of demand and you want to increase your power service just for a few moments. The best of all is that this increase of power could just be less than 1$, depending on your instance types and the time you’ll need to get the demand out.

What’s important to know about the cloud and this example? The cloud it is just interesting when your infrastructure could scale. And this is something that people doesn’t tell, or doesn’t understand very well. If my new instance could take a piece of the demand my graph is showing me it is because my software architecture has been developed to support this new instance.

In conclusion, cloud it’s great, in my opinion really great, but, and that’s very important, your architecture has to be prepared for new instances and you’ll have to manage your demand to get this new instance work. Once you’re sure your architecture can deal with this inconvenience, you’re prepared to the cloud!


Leave a Reply

Uso de cookies

Este sitio web utiliza cookies para que usted tenga la mejor experiencia de usuario. Si continúa navegando está dando su consentimiento para la aceptación de las mencionadas cookies y la aceptación de nuestra política de cookies, pinche el enlace para mayor información.plugin cookies