A Novel Efficient Query Strategy on Hibernate

Abstract

Database is an indispensable part of software development. In order to reduce unnecessary and redundant processes of accessing in database, the concept of Object-Relational Mapping (ORM) has been proposed and its corresponding applications have been widely accepted by developers. However, for a certain software system, accessing database systems is often one of the most common bottlenecks in performance improvement. In this paper, we propose the concept of "reference column(s)". In the actual business, there are many query operations with "reference column(s)" (non-primary) as parameters. Hibernate, as a popular ORM framework, provides excellent support for the above operation. Hibernate also implements a caching mechanism for storing query results. However, columns other than "reference column(s)" in these results may be frequently modified. Once any one of the above results has been updated, the corresponding searching could not be permitted to access the previous results storing in the cache. Accordingly, we propose a new hash-code based cache approach to avoid the emergence of the above problem. Extensive experiments on typical E-commerce datasets have been conducted to demonstrate the advantages of our approach.