Rdb Performance & Tuning: Physical Design
Length: 5 days
This seminar focuses on the ways you can improve the performance of your Oracle Rdb database application by manipulating the "physical" characteristics of your database and the system
memory allocated to the database. Importantly, the seminar also focuses on ways to avoid common pitfalls and known performance issues with the engine.
The seminar is divided into the following modules:
Context of Rdb You will examine the architecture of the product to understand what functions are performed by what processes and what the tradeoffs there are in the various run-time architectures. The goal is to help you understand what you can configure.
Transaction Management and Recovery You will learn how Rdb manages user processes bound to the database, where their transaction context is managed and why. Importantly, you will learn what parameters are potential limits to performance. Further, this module discusses database recovery both from a process and a node context.
Rdb Data Storage and Management - Part 1 You will learn how Rdb organizes disk storage and how Rdb stores data within that storage. This section discusses the logical partitioning of the database and addresses issues such as data compression. The primary focus of this module is uniform-format storage areas. An important component of this module is learning how to size storage areas and how and when to partition data.
Rdb Index Architectures You will learn about the various kinds of indexes that Rdb supports and how Rdb manages index structures. You will learn how to design physical index structures.
Rdb Data Storage and Management - Part 2 This module extends the material of the previous two modules to include mixed-format storage areas and both ordered and scattered hashed indexes. You will learn how and why to design hashed indexes and when to use and when not to use record placement within your database.
Rdb Memory Management - Part 1 This module addresses the ways that Rdb uses memory and shares access to database pages across processes and across VMS clusters. You will learn how to select global or local buffers for your database, how to size the buffer pool and how to tune your OpenVMS system for the use of global buffers. The materials include a spreadsheet that assists in some of the sizing decisions.
Rdb Memory Management - Part 2 Row caches represent the newest memory sharing technology. They enable Rdb to utilize ANY amount of memory available on a system for shared structures. Row caches have the potential for enormous performance advantages for certain database applications. This module addresses the reasons to use row caches and how to tune your database and system for row caches.
Dr. Jeffrey S. Jalbert, President