Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> Consider how Discord is begging you and everyone to sign up for Nitro because they're housing and serving all of their data. Most IRC networks on the other hand operate perfectly fine off of donated volunteer time and hardware for tens of thousands or even hundreds of thousands of users.

As of around a year ago Discord claimed to be passing around 4 billion messages a day.

I don't know how their average message size compares to Slack, but I just took a look at an export of messages from my company's Slack server and our busiest day in six years was just over 1MB in uncompressed JSON format, around 1.4KB per message. Compressed it was around 104 bytes per message. If we assume that the average message size is similar and similar amounts of metadata are stored per message, that means we're talking about somewhere between 400GB and 5.6TB per day for the entirety of Discord.

That's a lot of space on an individual basis, but nothing for a global-scale service. Obviously that's just for text and not any uploaded files, inline previews, thumbnails, etc. but still the point remains. Archiving text is not really a hard problem to solve. It's tiny by modern standards.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: