Use of Hybrid Mechanism of High Performance in Distributed Database Management System

Abstract- The frantic carrying out of operations in a distributed database management system has to meet the joint contingent properties of an ACID (especially, consistency and isolation) in order that the present simultaneous transactions can fail or pass in any way. Techniques such as Two-Phase Locking, Timestamp Ordering, Optimistic Concurrency Control, and Multi-Version Concurrency Control (MVCC) have something different with advantages and disadvantages under various high contention or skewed workloads. This paper proposes a hybrid concurrency control model, which can intelligently select among these protocols based on real-time transaction attributes and workload conditions. In case of another transaction being blocked by an active one, a transaction classifier and protocol allocator is set to switch back and forth dynamically between strategies. With the help of PostgreSQL and Python, simulating and evaluating the system using metrics such as throughput, latency, abort rate, and occurrence of deadlock performance. Overall, hybrid models can exhibit performance improvements, especially with read-intensive and mixed workloads, by - reducing transaction aborts and giving a faster response time. Hybrid models will offer modern scalable DDBMS a good alternative.