Revision as of 21:06, 23 June 2022

OSE Infrastructure

Check out [1]

Says 250 users per core
OSE is quad core - so we can handle 1000 concurrent database requests? Does that mean 1000 requests every second?
How long does a database request take?
If we can handle 1000 database requests - how many concurrent users is that? Would that make it 60k users simultaneously, because a request takes 1 second?
Optimal server cluster is 3 servers - [2]
wikipedia has 350 servers? [3]
Absolute max for a single server is 65k users connected, but really dependent on server specifications? How long is a connection - is it only during a database request? [4]
1000 users at a time is for 4 GB RAM? [5] Does that mean that we are likely to handle 16k simultaneous users?

@@ Line 11: / Line 11: @@
 *wikipedia has 350 servers? [https://en.wikipedia.org/wiki/Wikipedia:FAQ/Technical#:~:text=New%20servers%20bought%20as%20needed,number%20to%20about%20350%20servers.]
 *Absolute max for a single server is 65k users connected, but really dependent on server specifications? How long is a connection - is it only during a database request? [https://serverfault.com/questions/533611/how-do-high-traffic-sites-service-more-than-65535-tcp-connections#:~:text=On%20the%20TCP%20level%20the,65535%20simultaneous%20connections%20per%20client.]
+*1000 users at a time is for 4 GB RAM? [https://www.mediawiki.org/wiki/Topic:Vvi5md45il3wc9ho] Does that mean that we are likely to handle 16k simultaneous users?