Adaptive timestamp based concurrency control for efficient transaction processing in permissioned blockchains

Efficient concurrency control is key on permissioned blockchain platforms like Corda, where transaction volumes are high and contention is rife. In some hotspot situations, conventional concurrency control mechanisms such as Two-Phase Locking, Optimistic Concurrency Control, Multi-Version Concurrency Control, and Timestamp-Based Concurrency Control have high transaction abort rates, inefficient replay operations, and poor resource utilization. This paper presents an Adaptive Timestamp-Based Concurrency Control framework (ATBCC-Corda), which is based on four key components: Adaptive Conflict Prediction Engine (ACPE), Adaptive Hotspot-Aware Micro-Batching (AHMB), Adaptive Replay Selection Policy (ARSP), and Incremental Deterministic Replay (IDR), all of which are introduced to predict conflicts proactively, optimize the transaction scheduling, select the failed transactions for recovery selectively, and continuously adapt to the workload dynamics through an Adaptive Decision Manager (ADM). Further, a novel Blockchain Resource Efficiency Index (BREI) is proposed to jointly consider throughput, bandwidth consumption, and transaction abort rate. Experimental evaluation with BLOCKBENCH-inspired workloads shows that ATBCC-Corda consistently outcompetes Native Corda, Two-Phase Locking, Optimistic Concurrency Control, Multi-Version Concurrency Control, and conventional Timestamp-Based Concurrency Control in terms of throughput, abort rates, and width usage, and resource efficiency, making it a viable solution for scalable enterprise permissioned blockchain systems.