Serverless Use Cases
Serverless is the hot new technology. Everyone raves about it, and many have seen dramatic benefits from using it. Engineering overhead goes down, costs go down, fun goes up! But what is everyone actually doing with serverless technology?
We have been reaching out to the best engineers on the planet who are using serverless tech. Along the way we have learned a ton about where serverless tech works well, but also where there is untapped potential. Let’s take a look at what people have told us!
Many people using serverless find it’s efficiency of utilization powerful. Why run a server when you just need to do a few things every once in a while? Here are two example use cases we have heard many times:
- Cron jobs (e.g. AWS CloudWatch Events that trigger functions periodically)
- Storage bucket upload post-processing (e.g. S3 notification events)
In particular we have heard from people running personal side projects to companies running business critical processing as side-effects of S3 uploads. For example, people use S3 events that trigger Lambda functions to perform video transcoding, image manipulation, and public design contest submission analysis.
The most obvious use case to me, though maybe only because I came from companies with the need, is data ingestion pipelines. These pipelines take HTTP POST requests with data payloads, send them to an AWS Kinesis Stream, then process them using more Lambda functions as needed. Hooking up functions and streams creates much of the foundation for a Samza-like architecture without needing to figure out how to provision your own Hadoop and Kafka clusters.
Then there are the cases of crazy integrations everywhere. You can use a Lambda function to connect Facebook’s chat bots up to business logic that then sends data to your CRM. You might have an API service that stores data in AWS DynamoDB, which then emits events that can be processed as side-effects to do things like sending a follow-up email after a user performs an action. I have heard of a ton of use cases in this vein, but it also feels like there’s even greater untapped potential.
These are the biggest categories of serverless tech usage we have seen to date. But this is just a start. I believe we will see rapid expansion as people become more familiar with the tech and tooling, like Stackery, enables them.