Rationing data updates with consistency considerations in distributed systems

In this paper, we focus on reducing bandwidth utilization and processing cost by minimizing the amount of internal data updates in a distributed data system. The considered system can be a distributed database, a mobile application, or a communication protocol. Our essential approach is to rely on frequent internal updates instead of immediate change notifications to synchronize inter-dependent data units spread across a distributed system. Postponing updates would cause the recorded data to be inconsistent with reality. In our approach, inconsistency with reality is an option; however, data synchronization is a must. Users or external clients can either read consistent data after a limited delay in response time, or inconsistent data with no delay. Consistency with reality is probabilistically assessed to reflect the possibility of reading consistent or inconsistent information.