When you first learn cloud, you memorize a one-line formula: "Multi-AZ is high availability, Multi-Region is disaster recovery." But this formula states only the outcome, not the reason. Why does AWS bother placing multiple Availability Zones inside a single region, and why did it design the distance between those AZs to be "far enough, but not too far"? Hidden in the answer to that question is the most fundamental trade-off of distributed systems — you cannot optimize physical distance, data consistency, and fault isolation all at once.
An AZ is a cluster of one or more data centers with independent power, cooling, and networking