8 Comments

Nice summary! Determining beforehand which data is requested more often remains a challenging task. You have to monitor and adapt, yet it's still a very important component for many systems. CDNs are also evolving quite promisingly, with innovations like CloudFront functions allowing for latency-sensitive manipulations directly at the "edge."

Expand full comment

Thanks Steffen!

And I agree with your point about CDNs and the challenge to determine what data may be relevant. I feel data analytics can play a big role in this by tracking customer behavior and trying to come up with the most suitable items that should be pre-cached.

Expand full comment

I am just a beginner and got a better understanding of this concept. Thank you so much sir for such an amazing article!!!

Expand full comment

Thanks for the great feedback Divyam!

Happy to share about the concept

Expand full comment

I remember S3 had a big outage because they relied on the cache and once they lost caching servers in one region, they had a cascading failures of caching servers elsewhere. To recover the system they had to keep API down until the cache was preheated.

Big systems are often not capable to operate without cache.

Great post and thanks for mentioning Saurabh!

Expand full comment

Great point Eugene!

Cache is indeed quite an important component.

A lot of the high availability and performance-heavy systems wouldn't have reached such levels without the use of cache.

Expand full comment

Love this, especially with the analogy for the triple-edged sword!

10k subs is getting sooner and sooner, let's go!

Expand full comment

Thanks for the amazing feedback Daniel!

Expand full comment