What is Amazon EC2 – Part 3: instances

by Alexander Weiß

In this post I’ll explain what is an Amazon EC2 instance and what are the differences between them and virtual server.

In my last blog post I wrote about the features of Amazon EC2. The central element of Amazon EC2 is the instance. An instance is the working house in your Amazon cloud, it provides the computing power and everything else that you need to process data. Practically it is pretty much like a virtual server that hosting companies offer.  There are some major differences though: You pay for the instance only when the instance is running. If it runs nine hours a day, you pay for nine hours and not for the whole day. If you use an instance just for testing and would shut it down before you are leaving the office you would save quite some money. The concept of “pay only for what you use” is not used just for instances, though. Almost everything in the Amazon cloud works like that.

Instance types

There are a few things you have to decide when you want to launch an instance. First you have to decide which amount of the different resources (CPU, memory HDD) you need. Amazon offers 14 different instance types. You’ll find a full list here. The difference in the specs of the single instances is very big. The entry level is the micro instance with just a little bit more than 600 MB of memory and – for short bursts – up to two EC2 Compute Units.  The most capable instance “Cluster Compute Eight Extra Large Instance” comes with over 60 GB of memory and 88 EC2 Compute Units. There are also different instances for various use cases. For example, if you’ll have memory intense tasks running on your instance you can choose a high memory instance.

Reserved Instances

The benefit of “paying only for what you use” which I mentioned earlier in this post is only applicable for On-Demand Instances. If you already know that your instance will run 24/7 you can save some money by choosing a Reserved Instance. Here you pay a defined amount of money upfront for a discount on your hourly charge. For Light Utilization Reserved Instances you pay the least upfront but you also get the lowest discount. For Heavy Utilization Reserved Instances you pay the most upfront, but get the biggest discount.

There are also two time spans available: a 1 year and a 3 year plan. If you choose the 3 year plan, you have to pay more upfront but get a bigger discount.  Here the discount can go up to over 65%, but you have to pay the hourly costs for over two and half years upfront. But what happens if you have signed up for the 3 year plan but your business runs better than expected and the instance is too slow after a few months?

More flexibility for instances

Amazon has a neat solution for that: you can sell your EC2 instance in the Reserved Instance Marketplace.  However, this is not only a cost saver if you want to get rid of an instance, but also if you need more instances. Sometimes you might need an instance only for a short time or sometimes you might want to spend only a fixed amount of money. In these cases you should take a peek into the Reserved Instance Marketplace.

Another way to get cheap computing power is the Spot Instances. These instances allow users to make bids on Amazon EC2’s unused computing power. As long as the price of the instance is lower than your bid you can use the instance. However, if the capacity utilization of the Amazon EC2 raises than the price for computing power raises, too. And you can only operate the instance a s long as your bid is higher.

There is much more to say about instances, but I’ll stop for now, because I don’t want to write an overly lengthy post. But I’m pretty sure I’ll get back to that topic soon. In my next post I’ll show you how to get started with Amazon EC2.

VN:F [1.9.22_1171]
Rating: 0.0/10 (0 votes cast)
VN:F [1.9.22_1171]
Rating: 0 (from 0 votes)