Fan out / Fan in (Durable Functions)

A common serverless cloud pattern is using a fan out (scale workload horizontally) and then once done, fan back in, processing / finalizing the results.

Doing this in-process is fairly easy as you can see code in a "procedural" way. This simplifies progress, completion, error handling etc. but.... it doesn't scale.

Fanning out workload do various Functions, solves the scaling problem but introduces a lot of complexity.

Azure Durable Functions, allows you to scale out the workload to different functions, in a "procedural" way bringing together simplicity & scale

0:00 Introduction
0:23 Describing sentiment use case
1:20 Fan out / Fan in architecture
4:07 Looking at code
7:04 Orchestration constraints and conclusion