Lock Tracking – Efficient Solution for Relational Transaction Locking

Abstract

Relational database systems form an inevitable robust part of information technology. Data management is based on the relational schema. Individual operations are always encapsulated by the transactions. Transaction support ensures atomicity, consistency, isolation, and durability of the operations by shifting the database from a consistent data image to a new consistent image. Internally, individual operations can be divided to read and write operation sets. Generally, transactions are in a complex, scalable, and massive parallelism environment, where data rows are shared. The possibility of sharing without any operation loss must be ensured at any time. This paper deals with transaction locking and associates the problem of parallelism with indexing by proposing a new Lock Tracker module. It consists of the locked data rows perspective reference. By implementing such a solution, transaction hangs can be limited. Data access can be optimized, resulting in improved performance.