FaaS

FaaS is paid only for the user’s activity, i.e. for the amount of used storage space and number of operations during the month.

Working principle

Serverless computing is virtualization like other -aaS services, only it virtualizes the business logic of the services.

The customer does not pack the necessary piece of software code into a container, but moves it to a separate platform in the cloud. The customer can now run it when needed.

Features

Serverless computing provides a number of features:

1. Maximum flexibility. In standby mode, the service is not resource consuming, i.e. the customer does not load the system and does not pay for rent. In order to apply a function, the service allocates the required amount of resources.
2. Backend and Deployment. Development of a new function comes down to the easiest unloading of the code to the selected platform. The operation takes minimum time and effort, which helps developers to improve the software product faster and more efficiently.
3. Notifications. The notification system works in asynchronous mode, using different methods – SMS, e-mail, push.
4. Automation. Many additional services necessary for development are allocated in automatic mode (for example, code deployment, instance allocation or monitoring).
5. FaaS is a great option for Machine Learning, Internet-of-Things (IoT) and artificial intelligence (AI) development. This is due to product development on FaaS platforms which is done closer to the endpoint, i.e. users.

Amazon Lambda

The Amazon Lambda service is a great example. It performs all standard software code operations, has an integrated monitoring system, provides fault tolerance and can be scaled up.

The only exception is that the Lambda platform is a fully proprietary product. In other words, the customer is limited to all the features that Amazon has. In order to change or move to another vendor’s FaaS, the architecture will have to be rebuilt from scratch.