Capacity estimation

Capacity estimation is the rough arithmetic you do before designing a system, to work out how big it needs to be. You start from a few assumptions, like daily active users and how often each one does something, and turn them into requests per second, storage per year, and network bandwidth. Rounding hard is the point: a day has about 100,000 seconds, so 100 million requests a day is roughly 1,000 per second, and peak traffic is often a few times the average. The answers only need to be right to within a factor of ten or so, because their job is to tell you whether one database will cope or you need sharding, caching, or a CDN.