There is a massive shift toward serverless computing for big data and other applications on AWS Cloud. The popularity of serverless stems from the huge advantages it provides in terms of having no servers to provision or manage; it scales with usage and you never pay for the idle time. Additionally, availability and fault tolerance are built-in.

Serverless nicely fits into big data platforms as you can flexibly mix and match serverless, managed, and virtualized services. These services can be easily leveraged to rapidly ingest, categorize, and discover your data, allow easy query and analysis of your data, transform and load data, provide custom event-based handlers, and so on. Overall, a serverless approach allows you to focus more on analytics / use cases and not on infrastructure or servers.

A serverless strategy can be applied to almost all aspects of big data applications, including data processing, data warehousing, reporting, real-time processing, predictive analytics, and artificial intelligence.

Some examples of AWS serverless options include:

• AWS Lambda (serverless compute) lets you run your code in the cloud. It is a fully managed and highly available service that is triggered through an API or state changes in your setup. It scales automatically to match the incoming event rate. It provides support for code written in Node.js (JavaScript), Python, Java, and C#. It is currently charged per 100 ms execution time.
• Amazon Athena (serverless interactive query service) allows you to query directly from Amazon S3 using ANSI SQL. It supports multiple data formats and you pay per query.
• AWS Glue (serverless catalog and ETL/ELT service) helps crawl, discover, and organize data. It is supports integration with managed and serverless analytics, Job Authoring and Job Execution (serverless ETL). You pay for what you consume.
• Amazon Kinesis (serverless streaming) makes it easy to capture, deliver, and process streams on AWS.