A study for performance comparison of different in-memory databases

Abstract

In-memory databases have been researched for over three decades now. Since a major trend in OLAP systems, especially in Business Intelligence applications, is right and real time computing, in-memory databases may be an efficient solution to provide such requirements. There are several in-memory database systems on the market today. In several studies, the performance of a traditional disk resident database and a selected in-memory database is compared. However, there is a lack of a comparative study that evaluates different in-memory databases in the context of analytical applications. In this empirical study we compared performance of different in-memory databases. For this purpose, we selected Timesten, Altibase, solidDb and SQLite. Then we evaluated them both on Linux and Windows systems in terms of query response time using queries that need to access different amounts of data and are of different level of complexity. Our results show that, Altibase product gives the best results with Linux operating systems and TimesTen product performs much better than other in-memory products under the same conditions.