A Decade of Oracle Database Manageability

Oracle took on the challenge of ever-increasing management complexity for the database server beginning some eleven years ago with Oracle 10g and continuing through the 11g release. The Oracle RDBMS offers numerous services and features that share infrastructure and resources to enable a rich portfolio of industry-leading data management capabilities. Oracle Database Server Manageability refers to a cross-organization effort across various RDBMS core teams to deliver a coherent feature suite that enables customers to manage an Oracle database with consistency, confidence and relative ease through the Enterprise Manager UI. Our efforts to improve database manageability can be classified into three levels of increasing sophistication: • Statistics and reporting level– where key workload performance statistics and run-time data are continuously and efficiently collected by default, persisted into a self-managing repository, and exposed through reports and interactive UI • Advisory level – where instrumentation level performance data are periodically analyzed using a time-based methodology that makes findings and recommendations about sub-optimal performance, its root cause, and opportunities for improvement • Automated implementation level – where advisory level recommendations that can be implemented with low and well-understood risk are automatically put in place (given prior consent of administrators) Each layer builds upon the ones below it, and each plays a part in satisfying the key customer use cases – it is only through exposing features at all three levels that the solution becomes complete. In this paper we will discuss our key offerings at each level, review motivating use cases and lessons learned, and conclude by looking forward to some problems that excite us in the near future.