Performance Forecasting for Performance Critical Huge Databases

Fast databases are no longer nice-to-have --they are a necessity. Many modern applications are becoming performance critical. At the same time, the size of some databases has been increasing to levels that cannot be well supported by current technology. Performance engineering is now becoming a buzzword for database systems. At first physical and partially logical tuning methods have been used for support of high performance systems, but they are mainly based on large and not well understood performance and tuning parameters. Nowadays it becomes obvious that we need methods for systematic performance design. Performance engineering also means, however, support for database's daily operating. Most methods are reactive, i.e. they are using runtime information, e.g. performance monitoring techniques. It is then the operators or administrators business to find appropriate solutions. We target at active methods for performance improvement. One of the potential methods for active performance improvement is performance forecasting based on assumptions on future operating and on extrapolations for the current situation. This paper shows that conceptual performance tuning supersedes physical and logical performance tuning. As a proof of concept we applied our approach within a consolidation project for a databases-intensive infrastructure.