FALL CONFERENCE 2006

Hosted by PG&E

November 2, 2006

at PG&E, San Francisco, CA

See directions to the conference location near the bottom of this page.

Reflecting across nearly 20 years of solving Oracle performance problems, I can recognize a single pattern of behavior that is the dominant reason for failure in all the projects I've witnessed. In almost every case I've seen, failures in diagnosing and repairing performance problems have been caused by unrecognized SKEW in diagnostic data. This presentation shows several examples that illustrate why skew is such a pervasive problem for performance analysts.

In 2003, Cary Millsap and Jeff Holt prescribed a new Oracle performance problem diagnosis method called Method R in the book Optimizing Oracle Performance. The central tenets of Method R are that you should sequence your performance analysis tasks in business priority order, and that the most efficient way to analyze performance is to decompose the total end-user response time for those prioritized business tasks. Two years after its formal introduction, supporters of Method R cite extraordinary successes, and its critics describe various inadequacies. This session presents a refined understanding of Method R that incorporates some years of experience with the method since its invention. The session begins with a description of the method and its distinguishing features. It proceeds with a sequence of responses to commonly recurring questions that prospective Method R adopters have to answer.

Datapump is the next generation of Oracle's Import/Export utility. It's architecture is radically different offering many new benefits, as well as some serious gotchas for the uninitiated. This session explores the architecture and use of Datapump and includes numerous live demos of the new tool. Learn about performance and security enhancements, new features, significant differences from and the future of the old Import/Export, and critical methods for launching, monitoring, and stopping Datapump jobs. This session will be of benefit to anyone who is currently using the Import/Export utility, or who will be using Datapump in the future.

Real Application Cluster technology is gaining acceptance in the industry. It brings many benefits in terms of scalability, reliability, and high availability. It also represents some significant challenges to successfully install and operate. This session explores the critical success factors required to plan and install a RAC database, including both technology and people & process issues. Learn the requirements of storage and network engineering that must be addressed before even considering RAC, understand how to set realistic expectations from RAC, step through the decision making and installation process including new tools in 10gR2 and best practices. This session will be of benefit to anyone who is currently facing a RAC install, or in the planning stages of one.

Over the last several years, I've worked on a number of cases that involved performance, down servers, upgrades gone bad etc. Taking each major case in chronological order, what worked, what didn't, did we ever get root cause? Was 10046 trace, wait states, STATSPACK useful? Example -- user reports that Oracle was sorting dates incorrectly -- and the user was right? How can this be?

The session will cover the following topics in RAC:

• Quick RAC architecture review
• Determine RAC-specific tuning components
• Tune instance recovery in RAC
• Determine RAC-specific wait events, global enqueues, and system statistics
• Implement the most common RAC tuning tips
• Use the Automatic Workload Repository in RAC
• Use the Automatic Database Diagnostic Monitor in RAC
• Key Diagnostics

Although ultimate security may well be a "superstition," application security is an important component of system design. You must carefully design which users may access which parts of your application. This is true of all applications, but security plays an even larger role in web applications that are often open to access from a wider user base, sometimes even an Internet audience. The Oracle Application Server's component Oracle Containers for J2EE (OC4J) offers runtime services for Java-based web applications. OC4J provides a security feature, JAZN (Java authorization and authentication provider), which allows developers to use standard Java security libraries to implement user access and restriction features in their applications.

This presentation explains this feature of the OC4J server and discusses how it provides solid security services for J2EE web applications. The presentation explores different ways to store user credentials in JAZN and how to tap into existing Lightweight Directory Access Protocol (LDAP) providers such as Windows' Active Directory. It also describes the JAZN ability to access user credentials stored in XML files and how to switch back and forth between these two styles when moving from development to production credential stores. Then this presentation describes and demonstrates how to set up and code security hooks into your application using JDeveloper 10g so you can restrict and permit user access and operations to your valuable enterprise data.

Web application communications between client browser, application web server, and database server is usually a bit of a mystery to those accustomed to working with client/server and terminal mode applications. Any web application written in the Oracle environment works within web communications constraints and features. Knowing about the mechanisms involved is important for writing global application data; reading browser information; and debugging.

This presentation explains the mechanics of Hypertext Transfer Protocol (HTTP) -- the network protocol used for all web communications. It describes the roundtrip communication path between the client browser and application server; HTTP request and response messages; methods (get and post); the components of a Uniform Resource Locator (URL); and J2EE deployment files such as web.xml, server.xml, enterprise application archive (EAR), and web application archive (WAR) files. It also compares web communications to Forms code runtime.

Patterns in procedural programming became popular a decade ago, originated with a landmark book by Gamma et al. Each pattern has its name, so that developers could instantly refer to it and leverage it as a standard solution of a known problem. Unlike a majority of procedural programming languages, SQL is a high abstraction level language. Yet the advanced part of its learning curve demands a SQL programmer to develop the ability to recognize and communicate complex patterns. Beyond a certain point, the skill of piling up subqueries doesn't pay off, and one has to study some rudimentary theory, which classifies known SQL solutions into common design patterns.

Class models form the heart and soul of UML modeling, just as ER models form the core of Oracle method and information engineering. Poorly developed models result in brittle, low-quality systems. This presentation will present techniques and guidelines for avoiding common mistakes in class models, resulting in high-quality, robust models and systems.

Workload characterization is the process of identifying classes of workload, measuring those classes and then identifying their impact to the business. This presentation will explore how this can be done in an Oracle database including how to measure the impact workload classes have on database workload. This information will also be used to show how to enable an organization to better schedule its business processes.

This presentation walks you through how to create a basic Java/JSP Web Page that uses Oracle as a backend database providing data to fill a table. Discussion will focus primarily on Web Page basics like JSP Lifecycle, HTML/JavaScript, JSP Scriplet, JSTL and Database I/O with JDBC. However some fringe topics will be discussed like N-tier Architectures, and frameworks like Struts, JavaServer Faces, EJB.