Concept of Select Unlocking Optimization

Abstract

Ensuring data integrity stored in a database requires change encapsulation in a transaction. Transaction maintains consistency by shifting the database from one consistent state to another, protected by atomicity, isolation, and durability. This paper focuses on the isolation and parallelism allowing distributed access to the data tuple across multiple transactions. Access to the data tuple is marked by the locks applied on the row level. Locks can be shared expressing data retrieval process or exclusive intended to change the values of the columns. The general locking rule takes the lock before the operation and releases it just after the operation or at the end of the transaction. Oracle database, however, does not release the lock physically, whereas the data block access would be necessary to be reloaded during the transaction approval. Therefore, this paper aims in optimizing the process of unlocking data before the data retrieval process using multiple strategies.