Building a Serverless Web App with AWS
Breaking Down Amazon's Serverless Architecture & Services
by Shelly Lane
Going Serverless with AWS
The first lesson of going serverless with AWS is that you’re not truly going serverless in the literal sense. You’re resigning from the overhead and responsibility of provisioning, maintaining and administering servers and handing the responsibility for the AWS. It almost sounds too good to be true, right?
How Can I Move to a Serverless Architecture?
AWS offers a robust collection of products that can be used to replace the functions that servers are currently carrying out in your applications. A handful of these services will be detailed in this blog but there are plenty more to explore. ParkMyCloud provides a convenient list of services categorized by use case.
So what exactly is AWS Serverless? Using AWS Serverless means to build a serverless application using AWS’s fully-fledged army of services. This concept can seem a bit abstract at first. To provide a tangible example, let’s say that you’re thinking about building a serverless web application. The architecture below provides an example of how this can be accomplished using AWS services. AWS provides a good tutorial here that serves as a guide to creating your first AWS Serverless web application with this architecture.
AWS Serverless Services
The components shown in the example above are described in more detail below. Keep in mind that this architecture is only one example of how AWS services can be combined to build a serverless application.
Simple Storage Service(S3)
S3 provides a place to store data hosted by AWS. To get started, create a bucket which is a structure that serves as a container for related resources. Then upload your data to the bucket and choose a region for Amazon to store your data in. You can conveniently choose a region located near your end-users to reduce latency, if applicable.
Amazon Cognito
Amazon Cognito handles the user registration process and uses the user pool paradigm to manage user data and workflow. A user pool is a user directory that contains all user information. Amazon Cognito handles all processes related to users, including the implementation of a registration flow, sending registration emails, provisioning JWTs and more. Add two-factor authentication or SAML with the click of a button. These processes are fully customizable in the AWS Cognito interface.
Amazon API Gateway
Amazon API Gateway allows you to quickly create and publish APIs. API management and maintenance become simple since AWS takes care of most of the dirty work. APIs created with AWS Gateway can effortlessly process hundreds of thousands of concurrent API calls providing excellent reliability and capacity.
AWS Lambda
AWS Lambda allows you to write code that lives on servers managed by AWS. Lambda functions can be automatically triggered by other AWS services or can be called from a web or mobile app. For example, you could create an API using Amazon API Gateway which then triggers a Lambda function to execute a post action on a database. The use cases for AWS Lambda are endless and extend far beyond the scope of the web application architecture outlined above. Check out the Youtube playlist found here to see how big companies are using AWS Lambda.
Simplifying Development with AWS SAM
While AWS provides great interfaces for the implementation of each of their services, it’s much more convenient to be able to set services up and running from the command line. AWS SAM makes this possible. Define your serverless application by using a template that specifies all functions, API’s permissions, configurations, and events. Then use the AWS SAM CLI to package and deploy your application. The CLI also allows you to deploy and debug Lambda functions locally rather than having to use the AWS web interface for manipulation.
Is Going Serverless Worth It?
Quick! Let’s start transitioning all of our web applications to be serverless. Not so fast. Developing serverless applications will require shifts in development flow and time invested in gaining deeper understandings of AWS services. It may be overwhelming to take on the development of an entire serverless application at once. One of the positives of using AWS services is that they can be implemented independently. Transitioning just one API or implementing Amazon Cognitio could be a great first step to exploring what AWS Serverless has to offer. 
Almost all AWS services use an elastic cost model meaning that you only pay for what you use, similar to how you might pay for your water or electric utilities at home. Services scale in size to account for varying resource utilization over time. This model could prove to be very beneficial for applications or jobs that only use resources a few times a month. Rather than paying a monthly fee for a server, you pay for what you use, when you use it. Read more about the AWS service pricing model here.
The Bottom Line
The opportunities that AWS Serverless provides are endless and may alleviate a lot of the overhead that can result from provisioning, managing and maintaining servers. Since every application is different, it comes down to whether the elastic pricing and time saved by letting go of server management can be beneficial in your situation.