Optimizing Data Availability and Scalability with RP*-SD2DS Architecture for Distributed Systems

This work introduces Range Partitioning Scalable Distributed Two-Layer Data Structures (RP*-SD2DS), an innovative data storage architecture with the objective of enhancing data availability and scalability in distributed systems. By employing SD2DS and preorder-preserving RP*, this design avoids the need for a router or coordinator, ensuring dynamic adaptability. The main goal is to minimize system downtime by efficiently distributing data across two layers and increasing availability during partitioning operations in traditional SDDSs, thereby avoiding the bottlenecks associated with master- or coordinator-based systems. The proposed solution offers significant improvements over MongoDB, a well-known and robust system, with a single or three Mongos instances.