Reconfigurable Scalable State Machine Replication

Existing works on the scalability of State Machine Replication mostly rely on partitioning the application state in order to execute multiple requests simultaneously. These works either (i) do not allow for the system to grow dynamically, or (ii) do not support strong consistency for requests involving data in different partitions. In (i), the need for manual reconfiguration hinders the applicability of such systems to large scales. In (ii), the set of applications that can be built is severely limited. We propose a Reconfigurable Scalable State Machine Replication protocol without these limitations. We present and analyze the proposed algorithms.