How Discord Stores Billions of Messages (2017)

Yeah I think people forget that Skype actually owned the video game voiceserver market for a few years.

Ventrillo and Teamspeak and Mumble were all good. But you had to assign someone in your friend-group to manage the server. This meant paying for hosting to do it "the right way", and in turn one friend either paid the hosting themselves or you had to figure out how to split the cost. Then if someone else joined the group you had to split it with them, etc.. Some people would self-host teamspeak or ventrillo at their houses so you could avoid those costs, but now you are reliant on an unreliable system of one friend hosting your voiceserver on their desktop computer. This means that router mishaps could send it offline, them turning off their computer could send it offline, or if the teamspeak/vent daemon wasn't running then your whole server is offline.

Skype solved a lot of those problems because it was always online, no one had to manage a server, and it was free. It sucked in just about every other way as a game chat option, but the benefits of no-server-management, always-available, and no-cost, made an objectively inferior product dominate the world of game chat.

Discord simply took the features of teamspeak/mumble/ventrillo and combined it with the service benefits that skype offered. No more server cost sharing and no more server administration. But you still got the benefits of actual game chat servers like voice lobbies (as opposed to initiating calls like skype).

I really don't think Marketing is what made Discord successful. This is truly an example of someone who solved a need. We needed a product like teamspeak/ventrillo/mumble combined with a service like skype. Discord was that creation. It truly solved a problem for gamers. Gamers were not looking to cling to skype, but they were all using it. Discord created a product that fit into the market perfectly and the masses ran to it because the need was so big, and Discord solved the problem that gamers needed. The ease of setup also helped. Sending a single share link that someone simply clicked was all it took to join a server and start talking. I think that ease of setup is also an incredibly under-rated strength of Discord. In fact I would venture to guess that most gamers joined their first Discord server by clicking a discord share link that was sent to them via Skype.

I'm very much in the "use PostgreSQL unless you have absolutely proven to yourself that it won't work for your project" camp but in this case it really does look like moving to Cassandra was a good choice.

NoSQL scalable stores like Cassandra basically only work well if you have a very strong model of the queries that you will need to make.

In this case, that's exactly what they had: they knew what their read/write patterns looked like and they knew that they would be growing at hundreds of millions of rows per month, so easy horizontal scalability was a hard requirement.

The biggest weakness of classical relational databases like PostgreSQL come when you have a super high volumes of inserts (as opposed to updates) which will continue to grow your database over time, and you need to keep all of that data accessible for real-time queries.

They might have been able to achieve something like this using a PostgreSQL extension such as Citus, but it really does look like what they are doing fits Cassandra's sweet spot.

It's definitely improved over the years, but every remotely populated server I'm in uses bots for basic moderation features like ban words, proper bans/kicks (for example, temporary bans), warns, etc. There's still a long way to go in my opinion.

Also, RBAC which allows huge flexibility for users to run their community in whatever fashion they want.