Benoit Dageville

Self-Tuning for SQL Performance in Oracle Database 11g

Commercial database customers across the board list SQL performance tuning as one of the most time-consuming tasks for database administrators (DBAs). The 10g Oracle Database provides a feature called the SQL Tuning Advisor to simplify the task. The 11g release adds a new database feature, called Automatic SQL Tuning, that closes the feedback loop for the first time, fully automating the SQL tuning workflow and solving some SQL performance problems without any DBA intervention.

SQL memory management in Oracle9i

Complex database queries require the use of memory-intensive operators like sort and hash-join. Those operators need memory, also referred to as SQL memory, to process their input data. For example, a sort operator uses a work area to perform the in-memory sort of a set of rows. The amount of memory allocated by these operators greatly affects their performance. However, there is only a finite amount of memory available in the system, shared by all concurrent operators. The challenge for database systems is to design a fair and efficient strategy to manage this memory. Commercial database systems rely on database administrators (DBA) to supply an optimal setting for configuration parameters that are internally used to decide how much memory to allocate to a given database operator. However, database systems continue to be deployed in new areas, e.g, e-commerce, and the database applications are increasingly complex, e.g, to provide more functionality, and support more users. One important consequence is that the application workload is very hard, if not impossible, to predict. So, expecting a DBA to find an optimal value for memory configuration parameters is not realistic. The values can only be optimal for a limited period of time while the workload is within the assumed range. Ideally, the optimal value should adapt in response to variations in the application workload. Several research projects addressed this problem in the past, but very few commercial systems proposed a comprehensive solution to managing memory used by SQL operators in a database application with a variable workload. This paper presents a new model used in Oracle9i to manage memory for database operators. This approach is automatic, adaptive and robust. We will present the architecture of the memory manager, the internal algorithms, and a performance study showing its superiority.

Oracle database replay

This paper presents Oracle Database Replay, a novel approach to testing changes to the relational database management system component of an information system (software upgrades, hardware changes etc). Database Replay makes it possible to subject a test system to a real production system workload, which helps identify all potential problems before implementing the planned changes on the production system. Any interesting workload period of a production database system can be captured with minimal overhead. The captured workload can be used to drive a test system while maintaining the concurrency and load characteristics of the real production workload. Therefore, the test results using database replay can provide very high assurance in determining the impact of changes to a production system before applying these changes. This paper presents the architecture of Database Replay as well as experimental results that demonstrate its usefulness as testing methodology.

Efficient and scalable statistics gathering for large databases in Oracle 11g

Large tables are often decomposed into smaller pieces called partitions in order to improve query performance and ease the data management. Query optimizers rely on both the statistics of the entire table and the statistics of the individual partitions to select a good execution plan for a SQL statement. In Oracle 10g, we scan the entire table twice, one pass for gathering the table level statistics and the other pass for gathering the partition level statistics. A consequence of this gathering method is that, when the data in some partitions change, not only do we need to scan the changed partitions to gather the partition level statistics, but also we have to scan the entire table again to gather the table level statistics. Oracle 11g adopts a one-pass distinct sampling based method which can accurately derive the table level statistics from the partition level statistics. When data change, Oracle only re-gathers the statistics for the changed partitions and then derives the table level statistics without touching the unchanged partitions. To the best of our knowledge, although the one-pass distinct sampling has been researched in academia for some years, Oracle is the first commercial database that implements the technique. We have performed extensive experiments on both benchmark data and real customer data. Our experiments illustrate the this new method is highly accurate and has significantly better performance than the old method used in Oracle 10g.

Parallel SQL execution in Oracle 10g

This paper describes the new architecture and optimizations for parallel SQL execution in the Oracle 10g database. Based on the fundamental shared-disk architecture underpinning Oracles parallel SQL execution engine since Oracle7, we show in this paper how Oracles engine responds to the challenges of performing in new grid-computing environments. This is made possible by using advanced optimization techniques, which enable Oracle to exploit data and system architecture dynamically without being constrained by them. We show how we have evolved and re-architected our engine in Oracle 10g to make it more efficient and manageable by using a single global parallel plan model.

Self-managing technology in database management systems

This chapter introduces self-managing technology in database management systems. All major providers of information technology have acknowledged the importance of the total cost of ownership (TCO) problem and have attempted to rectify it in one way or another by making their systems more self-managing or “autonomic.” While the idea of self-managing is easy to articulate, it is extremely challenging to design and build a completely self-managing information system. This chapter introduces the core concepts in self-managing technology. It focuses on providing self-manageability for relational database management systems, and how recent releases of IBM DB2, Microsoft SQL Server, and Oracle have embedded tools and techniques to enhance self-manageability. The core concepts and examples of self-managing database technology are useful for researchers who are interested in contributing to the area of reducing total cost of ownership for databases.