https://specbranch.com/posts/one-big-server/
Idea
The cloud era began in earnest around 2010. At the time, the state of the art CPU was an 8-core Intel Nehalem CPU. Each core offered 4 FLOPs per cycle, meaning that your 8-core server running at 2.5 GHz offered a blazing fast 80 GFLOPs.
The boom in distributed computing rode on this wave. Everybody had clusters of servers because one machine was not enough or because it was fashionable.
But times have changed. Today one big server is enough to host a web service with up to 10k QPS (queries per second).
The Server
- 64 cores
- 128 threads
- ~2-2.5 GHz clock
- Cores capable of 4-6 instructions per clock cycle
- 256 MB of L3 cache
this server is capable of 4 TFLOPs of peak double precision computing performance. This server would sit at the top of the top500 supercomputer list in early 2000. It would take until 2007 for this server to leave the top500 list.
It can serve 500k requests per second with nginx: https://openbenchmarking.org/test/pts/nginx
You can rent a similar machine for 1000-1500 €/month
Availability
An additional Backup Server should be enough. But you could use even a second Datacenter (2x2 Servers).
But beware of correlated hardware failures.
E.g. rent a backup server of a different type.
Inventor
Nima Badizadegan