Performance measurement tools are very important, both for designers and users of Database Management Systems (DBMSs). Performance evaluation is useful to designers to determine elements of architecture, and, more generally, to validate or refute hypotheses regarding the actual behavior of a DBMS. Thus, performance evaluation is an essential component in the development process of well-designed and efficient systems. Users may also employ performance evaluation, either to compare the efficiency of different technologies before selecting a DBMS, or to tune a system. Performance evaluation by experimentation on a real system is generally referred to as benchmarking. It consists of performing a series of tests on a given DBMS to estimate its performance in a given setting. Typically, a benchmark is constituted of two main elements: a database model (conceptual schema and extension), and a workload model (set of read and write operations) to apply on this database, following a predefined protocol. Most benchmarks also include a set of simple or composite performance metrics such as response time, throughput, number of input/output, disk or memory usage, and so forth. The aim of this article is to present an overview of the major families of state-of-the-art database benchmarks, namely, relational benchmarks, object and object-relational benchmarks, XML benchmarks, and decision-support benchmarks; and to discuss the issues, tradeoffs, and future trends in database benchmarking. We particularly focus on XML and decision-support benchmarks, which are currently the most innovative tools that are developed in this area.
Data-Centric Benchmarking
