What is Serverless?

How AWS serverless works + pros and cons

In this article, we want to have a quick overview of AWS Serverless and describe the pros and cons. Later we will publish our own case studies and explain how we implement Serverless Architecture for our clients.

Published date: 05-11-2021

1. Traditional Approach:

Develop software, build infra, scale infra, and software. It relies on having product development engineers, operations, and DevOps engineers, all of them working together to make sure a product is delivered on time.

2. What is the problem?

Business cares about time to market and agility. Operations and cloud costs are easy to measure, but the development time required to release a product to market is hard to estimate, forecast and plan. Therefore, forecasting demand is hard.

Total Cost of Software development Diagram
Figure 1

3. How Serverless can help?

The term serverless itself is baffling as most new people to it might judge it as being a practice without any servers to host and run the code. Alternatively, it refers to the approach in which end-user exclusively focus on their business tasks and model instead of spending time and money on server status, maintenance, scaling, capacity, and OS patching. All in all, it is defined as the practice of building a running application without managing servers. Even though the cloud vendors still must manage servers, VMs, containers, and Operating system that is abstracted from the developer

4. How does Serverless work?

Evidently, serverless is the next logical path on the advancement of cloud computing. Starting with bare metal servers following with the virtual machines and containers(Docker), the computing had accomplished to the era of serverless where one does not crave about the server and its maintenance anymore.

AWS has a mature set of services to build an application on its serverless platform. The idea behind serverless is that the customer only writes the business logic and leaves everything else to AWS (significant reduction in the cost of development). The core concept is pretty much clean where Serverless (like Amazon Lambda) runs developers’ codes in response to specified events, scales automatically, and provides built-in Amazon watch code monitoring and logging. Every Lambda function created contains the code to execute, configuration to mention how the code will be deployed, and event sources to observe events and invoke the functions

AWS Lambda: Users would only pay for the active functions, while stand-by functions only use sources if needed
Figure 2: Users would only pay for the active functions, while stand-by functions only use sources if needed

5. What services are available in the AWS Serverless Platform?

Computing: AWS Fargate, AWS Lambda
Messaging: Amazon SQS, Amazon SNS
Integration: Amazon API Gateway, AWS Step Functions
Security: Amazon Cognito, AWS Key Management Services
Monitoring: Amazon Cloudwatch, Amazon Quicksight
Database: Amazon DynamoDB, Amazon Aurora
Analytics: Amazon Kinesis, Amazon Athena

Subscribe