Yea I was thinking that there is some sort of parallel to datacenters, where it's lowest latency to route everyone to their closest DC but in practice people usually route a certain percentage of traffic to standby datacenters which are farther away to make sure they're still working.
You lose some efficiency but gain in resilience. Nowadays big internet companies have all sorts of Disaster Recovery plans.
BCP is the way, this is all risk management. Get financial lenders to insist on insurance, the insurance companies will figure out how to measure supply chain risk.
> where it's lowest latency to route everyone to their closest DC but in practice people usually route a certain percentage of traffic to standby datacenters which are farther away to make sure they're still working.
Are you keeping your own data center in addition to the AWS/GCP/Azure? Because that's the exact same premise.
