Hacker News new | ask | show | jobs
by martinald 1101 days ago
But you're missing the point above. S3 is probably the easiest service to replace - there are loads of providers which use the _exact_ same protocol as S3. It's a drop in replacement. It literally uses AWS concepts, there is nothing else to learn apart from putting a different url into your application.

Very few people should really be using S3 at any serious scale is my thoughts. The cost savings are enormous (plus cloudflare for example replicates your data a lot closer to users for no extra cost, significantly improving performance). The cost savings can be absolutely enormous for very little/no additional complexity given how many providers are compatible with S3, and the fairly 'boring' nature of S3 compared to other technologies.

1 comments

People really have to be afraid of running VPS or whatever if they can't spun up a min.io instance to have their own S3 without dealing with Amazon's Bullshit
How big of a production deployment have you ran on min.io in terms of data transfer / month and total storage?

Because I've done small scale and can tell you I'd run S3 in the future.

Point taken as I haven't used it in prod properly, I'll cross that bridge when I find it but Id rather keep my money and put some time into making it work. There's also cloudflare R2 now if S3 is too expensive but self-hosted is out of the question