Introduction to AWS

Let's move on to the first official section of this chapter, which gives you an introduction to AWS. In this section, we are going to take a look at Lambda usage and pay-per-use pricing, and also where to find documentation and other developer resources:

1. Let's go to the home page for aws.Amazon.com/lambda, as shown in the following screenshot:

2. Head over to the Documentation page. The Documentation page gives you links to many useful resources, such as SDKs and tools. But, for now, let's take a look at the Developer Guide.
3. The Developer Guide gives us a lot of useful background information on how Lambda works.

4. Click on the section called Building Applications with AWS and click on the Event Source Mapping:

5. Scroll down a bit and you will be able to see an example of how we can use Lambda, shown in the following screenshot:

In this example, Amazon S3 pushes events and invokes a Lambda function. Amazon S3 (or Amazon Simple Storage Service) is a scalable web service for storing and retrieving large amounts of data. In this example, we have a user who uploads a file into an Amazon S3 bucket. This triggers an object-created event. The object-created event is detected by Amazon S3, which triggers S3 to invoke our Lambda function. The Lambda function is associated with an execution role. The execution role gives our set certain permissions. So, in this scenario, Amazon S3 needs to have permissions to invoke our Lambda function, otherwise any other service would be able to invoke our Lambda function, which we want to avoid. So, if these permissions are given, our Lambda function is invoked with the event data from our Amazon S3 service invocation. This is also referred to as the push event model, but there's another way to use AWS Lambda. Let's scroll down a little bit to the next example:

Here, we have a stream-based service. In this example, Lambda pulls events from an Amazon Kinesis stream and invokes the Lambda function. On the left-hand side, you can see a custom application that writes data on a Kinesis stream. On the right-hand side, our Lambda function continuously picks up pieces or records from this stream. Again, we have an execution role associated with our Lambda function, but in this case it works the other way around. In this case, we need to give our Lambda function permission to access the Kinesis stream because here we are in the so-called pull event model. Whenever we pick up a new record, the Lambda function is executed.

Cloud service pricing

Now let's take a quick look at cloud service pricing. Cloud services work quite differently from traditional web hosting, and this also applies to pricing. With the traditional web hoster, you typically sign up for a long-term contract, maybe one year or two years, and you pay for the resources whether you use them or not. With cloud services, this works quite differently. With cloud services, you only pay for the resources that you actually use with very fine granularity. The downside of this is that the pricing model becomes a bit more complicated:

As you can see in the preceding screenshot, each Amazon web service has its individual pricing model. Typically, it breaks down into charges for the compute capacity, storage capacity, and data transfer that you use.

Let's take a closer look at the pricing model of AWS Lambda. The pricing model breaks down into two parts:

• First, you pay for requests, which is actually quite cheap. It's only 20 cents for 1 million requests.
• The other thing that you pay for is duration, which is the time that your Lambda function runs for. This time period is rounded up to the nearest 100 milliseconds. So, if you have a short-running Lambda function that only runs for 50 milliseconds, you pay for 100. If you have a Lambda function that runs for 910 milliseconds, you pay for 1 full second. You also have to pay for the amount of memory that you allocate to your function. You can configure your Lambda function with different levels of memory. You then get charged this fixed price, price constant, for every gigabyte-second that you use.

Let's take a quick look at a sample calculation. When you scroll further down in the page, you will see different pricing examples. Let's have a look at the first pricing example.

Pricing example

In this example, you will configure your Lambda function with half a gigabyte of memory. We will assume that the Lambda function is called 3 million times within one month. Each Lambda function, we will assume, runs for one second. With this in mind, our monthly compute charges would be calculated as follows:

Given these details, we need to calculate the total amount of time that our Lambda function is running for: 3 million invocations X one second per invocation is equal to 3 million seconds. Then we calculate the compute capacity that is used during these invocations. We use Lambda functions for 3 million seconds, and each Lambda function is allocated with half a gigabyte of memory, so we use 1.5 gigabyte-seconds. However, Lambda comes with a free tier, so up to a certain level you get compute capacity and requests for free. So if you deduct these from your calculation, then you end up with 1.1 gigabyte-seconds. To calculate this, you multiply that with your fixed price constant and you end up with roughly 18 dollars per month:

You must also pay for request charges. However, this only costs 20 cents per million requests, and the first million requests are free, so you only have to pay for 2 million requests, which in other words will cost you only 40 cents.

So your final calculation for the monthly charges will amount to roughly 18-19 dollars per month (both the compute and request charges).

The next section is about the AWS web dashboard, the so-called Management Console. So let's dive into that!