<![CDATA[This review examines resilient data analytics pipelines as critical infrastructure for fault-tolerant smart manufacturing systems. It addresses the need for reliable, low-latency, and integrity-preserving data flows in industrial environments where sensor failures, network disruptions, data corruption, platform faults, and cyber-physical disturbances can compromise real-time analytics and autonomous decision-making. The study synthesizes recent literature on Industrial Internet of Things-enabled manufacturing, distributed stream processing, edge-fog-cloud computing, fault-tolerant architectures, and data governance. It analyses pipeline layers covering data sources, edge preprocessing, fault-tolerant ingestion, stream analytics, distributed storage, security, governance, and decision-support applications. Core resilience mechanisms examined include redundancy, replication, monitoring, checkpointing, rollback recovery, graceful degradation, secure aggregation, audit logging, and adaptive recovery. The review also evaluates technologies such as Apache Kafka, Apache Flink, Apache Spark, Storm, HDFS, Cassandra, MongoDB, cloud-agnostic platforms, microservices, and digital twins. Findings show that resilient analytics pipelines support predictive maintenance, real-time process monitoring, quality assurance, supply-chain optimization, and autonomous manufacturing by preserving data continuity and analytical reliability during faults. However, major challenges remain in balancing scalability with performance, maintaining data integrity across heterogeneous edge-fog-cloud layers, achieving low-latency recovery, integrating legacy systems, securing distributed data flows, and establishing standardized design frameworks. The review identifies future directions in AI-driven fault detection, digital-twin-based resilience testing, scalable distributed architectures, autonomous data management, and benchmarking protocols. Resilient pipelines are therefore essential operational assets for sustaining reliable, secure, and fault-tolerant smart manufacturing intelligence.]]>
Copyrights © 2026
