Scalable Distributed Architectures for Real-Time Data Processing: A Novel Approach to Adaptive Analytical Querying

Real-time analytics demands scalable distributed architectures that can balance performance and consistency. This work presents R-Store, a novel integration-driven architecture combining adaptive query execution, stream processing, and hybrid OLAP-OLTP capabilities. Evaluated on a 144-node testbed using a Zipf-distributed TPC-H workload, R-Store achieves over 100K updates/sec with analytical accuracy and timestamp-consistent cube views. It outperforms traditional streaming systems by 27% in throughput and demonstrates efficient cube maintenance and query execution with predictable I/O cost modeling. Our architecture contributes a reproducible, low-latency solution for next-generation real-time analytics.