One of the biggest factors powering the nonrelational database revolution is a desire to escape the restrictions of strict ACID consistency. It’s widely believed that the new breed of nonrelational databases provide only weak or at best eventual consistency, and that the underlying consistency mechanisms are simplistic. This belief represents a fundamental misunderstanding of nonrelational database systems. Nonrelational systems offer a range of consistency guarantees, including strict consistency, albeit at the single-object level. And in fact, there are some complex architectures required to balance an acceptable degree of consistency when we lose the strict and predictable rules provided by the ACID transaction model.