Streaming Byzantine-Resilient Storage Architecture for Sensor Data in Intercloud Environments

Abstract

Sensor network growth in areas like critical infrastructure, smart cities, and environmental monitoring has resulted in previously unheard-of amounts of real-time data that need to be stored in a safe, dependable, and scalable manner.  Traditional cloud storage solutions frequently fall short because of their centralized failure points, high streaming data latency, and poor fault tolerance. This paper proposes Streaming DepSky-A, a novel extension of the Byzantine fault-tolerant DepSky protocol that supports real-time, block-based sensor data ingestion across multiple cloud providers. Unlike static file-oriented systems, Streaming DepSky-A operates on streaming blocks, enabling low-latency ingestion, efficient memory usage, and fine-grained fault isolation. The system integrates quorum-based replication, block-level integrity verification, and asynchronous dispatch to achieve high throughput and resilience under concurrent workloads and simulated cloud faults. Experimental evaluation in a federated cloud testbed demonstrates that Streaming DepSky-A sustains throughput exceeding 480,000 records per second per node, maintains over 99.94% availability, and detects data corruption with 100% accuracy. The results affirm the viability of the architecture for scalable, fault-resilient sensor data storage in untrusted and heterogeneous intercloud environments.