We present a database replication architecture that guarantees ACID transaction properties as well as high throughput expected of modern database systems. Higher throughput results due to server replicas processing distinct, non-overlapping subsets …