I would like to make this series of [TIL] as my knowledge base, so basically I will make this as my library. Maybe could be useful for you too.
Build a system (example : API for regional data) which can handle 1000 requests per second. On the second month requests increase to 10,000 per second, then third month increase to 100,000 per month.
Vertical Scaling :
Increase current server : make higher RAM, CPU, HDD.
Build new server with same specifications
Comparison between Vertical and Horizontal Scaling :
- Fault Tolerance
VS : need downtime if something happens on server
HS : no need to downtime, since another servers still works
- Low Latency
HS : distributed area, will choose the nearest server. So it will reduce latency response time.