Balancing consistency and performance in edge-cloud transaction management

Abstract

The proliferation of Internet of Things (IoT) devices has led to edge-cloud computing paradigms where resource-constrained edge devices connect to cloud servers. However, traditional concurrency control methods like two-phase locking (2 PL) and optimistic concurrency control (OCC) are inefficient in these heterogeneous environments. This paper presents adaptive transaction management protocols for edge-cloud systems. We propose EC-Lock which transitions between non-blocking and blocking phases, and EC-OCC which distinguishes edge and cloud transactions during timestamp validation. These hybrid techniques reduce unnecessary blocking and restarts. Simulation studies demonstrate that EC-Lock and EC-OCC provide substantial performance gains over traditional protocols under diverse workloads. By balancing consistency and efficiency, the proposed protocols enable scalable edge-cloud transaction processing. Our results show EC-Lock and EC-OCC better utilize scarce edge resources while minimizing cloud transaction impact. This work delivers innovative adaptive concurrency control optimized for emerging IoT-based edge-cloud computing architectures.