3 ms·

I think you're overall point is correct, but the specifics about twitter are backwards. Twitter would have been way better off using Cassandra (but it wasn't b

I think you're overall point is correct, but the specifics about twitter are backwards. Twitter would have been way better off using Cassandra (but it wasn't built yet!), instead you're 100% right, they did their own bespoke stack, trying to replicate what google had internally, and they didn't have the $$$ to do it. It was Facebook/Meta that later open sourced Cassandra and a buncha other great open source stuff.

jmyeet 7mo ago
Someone else suggested I was misremembering the Twitter Cassandra timeline so I went back and checked. No, Twitter droped MySQL for Cassandra in 2010 [1], after Facebook open-sourced it in 2008 [2]. Twitter claimed they were using Cassandra (or at least planned to) for storing tweets [3] but had rolled something else entirely (called Manhattan) by 2014 [4][5]. So yes it was originally released by Facebook but it was Twitter who spent a massive effort trying to make it work in production. And failed. [1]: https://www.informationweek.com/it-infrastructure/twitter-drops-mysql-for-cassandra https://www.informationweek.com/it-infrastructure/twitter-dr... [2]: https://en.wikipedia.org/wiki/Apache_Cassandra https://en.wikipedia.org/wiki/Apache_Cassandra [3]: https://highscalability.com/so-why-is-twitter-really-not-using-cassandra-to-store-tweets/ https://highscalability.com/so-why-is-twitter-really-not-usi... [4]: https://blog.x.com/engineering/en_us/a/2014/manhattan-our-real-time-multi-tenant-distributed-database-for-twitter-scale https://blog.x.com/engineering/en_us/a/2014/manhattan-our-re... ]5]: https://news.ycombinator.com/item?id=7515995 https://news.ycombinator.com/item?id=7515995