Give a brief description what we understand under the Read-After-Write Problem and how we're able to solve it.

When writing to one replica and reading from another, clients may receive a different value than they have written
 


When solved with writing/reading from both replicas, the fault probability increases
 
We can solve this by using three replicas
Borh reads and writes are sent to all replicas
A successful request requires acknowledgment of a majority (2 out of 3) of the replicas

Diskussion