System Design Interview Prep: Queue vs PubSub

This video contains a quick analysis of the difference between queue based architectures and publish/subscribe (pub/sub) architectures.

Although I've been using both queues and pubsub brokers throughout my career to handle big data workloads that process tens of terabytes of data every day, I admit reluctantly that I incorrectly thought queues were pubsub mechanisms.

It's easy to see why I, and likely many other Software Engineers, fail to make the distinction between queues and brokers - both are instances of asynchronous messaging mechanisms, and popular tools like Kafka (or AWS SQS + SNS) blur the lines between the two in providing very useful features. Asynchronous messaging is very useful when it comes to handling big data systems.

The video is effectively describes a Venn diagram that highlights the benefits of queues over brokers (such as parallel processing) and the benefits of brokers over queues (such as easy system extension).

Defining concepts precisely will help you solve more problems, whether that's a problem at work, an error at work, or even (sigh) a tech interview.

Chapters:
00:00 - Queue != Pub/Sub
00:42 - Why They Are SIMILAR
01:19 - Pub/Sub vs Queue
02:12 - Queue vs Pub/Sub
02:42 - Summary

References:

[siddharthgupta]

Why does this have so less views? Good and crisp explanation. Exactly what I was looking for! Thanks!