NoCOUG Conference 100

Sponsored by Quest Software

Wednesday, November 9, 2011

Computer History Museum

Mountain View, CA

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

“Coding Therapy for Database Professionals” - Steven Feuerstein, Quest Software

All database professionals must be able to write good code, whether it is SQL code, PL/SQL code, Java code, or something else. But, we can't write code without our brains, and our brains come with a full load of issues. The way our brain remembers the past and projects into the future has a big impact on how we write code. Moving beyond physiology, human psychology also plays its role, making it difficult for us to acknowledge ignorance and ask for help. In this keynote address, Steven Feuerstein will offer an intensive coding therapy session (including couples therapy, dream therapy, and shock therapy) to help all attendees come to grips with their innate, unavoidable issues and making it easier to write better code - and help others on their team write better code.

An idea being widely propounded among many software developers is that software should be allowed to freely access data without any regard for how or where the data is stored. Although this idea sounds appealing, it frequently fails in practice. Software applications that treat data storage as a black box often suffer from widespread problems with performance, contention issues, and high resource consumption. These problems can severely diminish the effectiveness of multithreading, added hardware, and other techniques used to support high application throughput and low response times.

This presentation will examine several areas where a lack of understanding of Oracle has repeatedly led to costly design flaws that are directly responsible for performance problems. The topics will cover commits and rollbacks, common mistakes made with unique identifiers, use of DBMS_LOCK by application code, and why you should filter data in the database instead of an application. The intent of the presentation is to help both developers and DBAs understand, recognize, and avoid several common pitfalls arising from the black-box approach to data.

“The History of Oracle Performance Analysis” - Craig Shallahamer, Orapub

This presentation will take an interesting (insightful, humorous, and just plain fun) stroll through Oracle performance analysis, starting from the beginning. It will reference books, methods, products, stories, and timelines - and even try to dig up some old spreadsheets full of ratio analysis work.

The Oracle Database Appliance is a new way to take advantage of the world's most popular database in a single, easy-to-deploy and -manage system. It's a complete package of software, server, storage, and network that's engineered for simplicity. Learn how you can save time and money by simplifying deployment, maintenance, and support of database workloads, and quickly scale from two processor cores to 24 processor cores without incurring the costs and downtime usually associated with hardware upgrades.

“Under The Hood of Oracle ASM: Fault Tolerance” - Alex Gorbachev, Pythian

Oracle Automatic Storage Management (ASM) has introduced a new concept of mirroring that is implemented differently than in any known RAID levels. So what happens when not one but two or more disks fail? Is such a situation hypothetical and highly unlikely? This session will help attendees to evaluate the data loss risks and adopt the best ASM configuration according to their risk profile. For a better understanding of ASM reliability features, this presentation will peek under the hood of ASM and provide live demos simulating ASM disk failures and ASM handling of such failures.

“Making the Most of PL/SQL Error Management Features” - Steven Feuerstein, Quest Software

Every well-crafted application follows a standard: a consistent process for handling, raising, and reporting errors. PL/SQL offers a powerful, flexible exception-handling architecture, but it does have some limitations. The presentation will show you how to take advantage of the latest error-management features, work around limitations, and rely on generic, reusable programs to raise, handle, and log exceptions. Go beyond the basics of exception handling in PL/SQL to explore the abundance of specialized error-management features in Oracle, plus receive recommendations for achieving consistent, robust application construction. After attending, you'll be able to more fully leverage PL/SQL error-management features and ensure that best practices for consistent error management are followed. There is so much to learn and there are so many ways to improve our code! Discover them here.

The SQL Monitoring Report was introduced in Oracle Database 11g and has become the single most used diagnostic report by the Oracle Real-World Performance Group and Oracle database development for SQL statement performance issues. This session will start with a technical overview of the SQL Monitoring Report and what metrics and information it provides. From there we'll deep dive into examples from the field explaining how this report was used to diagnose and validate performance issues. There is a wealth of information in the SQL Monitoring report, and this session will provide the necessary knowledge to best leverage it.

“Resolving Buffer Busy Waits” - Craig Shallahamer, OraPub

Of all the Oracle wait events, the buffer busy wait and read by other session are two of the most intriguing. While Oracle never officially locks a buffer, during high concurrency situations buffers can become too busy to access. A methodical step-by-step diagnosis and a solid grasp of the relevant Oracle internals is particularly important because there are a number of reasons why a buffer can become too busy, and the solutions are derived in part by understanding the underlying Oracle structures. This presentation guides you through the process of discovering if there is a significant problem, and the related Oracle internal structures, wait event name changes, a spot-on diagnosis method, and solutions for the most common situations you're likely to face. The presentation closes with a case study bringing together all the presented material.

“Visual SQL Tuning By Example” - Kyle Hailey, Delphix

Database load is created with SQL; thus, it makes sense that performance bottlenecks are most often caused by poorly performing SQL statements. For analyzing poorly performing SQL statements, this presentation will show you a solid step-by-step method for analyzing, understanding, and tuning the problem SQL though Visual SQL Tuning (VST) diagrams.

VST is a method of laying out the tables and joins of a query graphically, indicating key features of the query in the graphics. Through the VST, you'll learn how to quickly visualize any coding errors in the query, discover flaws in the underlying database schema, and most importantly, find the best execution path through the query.

Oracle recently announced the Oracle Public Cloud that runs on Oracle Engineered Systems, providing customers and partners with a high-performance, reliable, elastic, and secure infrastructure for their critical business applications. The Oracle Database Cloud Service provides access to Oracle Database 11g, application development tools, data loading services, Web Services APIs, and a set of packaged business productivity applications. This session will demonstrate how to get started and how easy it is to start deploying and developing applications.

Oracle's cost-based optimizer, for all its years of intelligent design, has not ended the need for occasional manual SQL tuning. Manual SQL tuning should begin with the process of manually working out what is the best possible path to the data called for by a SQL statement (usually a query), while simultaneously checking that the SQL avoids the many pitfalls that are usually characteristics of mis-coded SQL that fails even to make good sense. The path found by manual SQL tuning may turn out to be the path already chosen by the optimizer, or may turn out to be simply unavailable to the optimizer for good technical reasons, or may be one that the real optimizer, or even a hypothetical perfect optimizer, could choose, but does not, at least not yet, for technical reasons. I will explain why this manual SQL tuning process is a useful and efficient means to identifying and solving the underlying technical problem even in many cases where the optimizer is doing its job perfectly, and therefore will remain a necessary process for some SQL even if the optimizer were hypothetically to achieve utter perfection.

During the optimization of a SQL statement, the Oracle database optimizer relies heavily on statistics to estimate the number of rows produced by each of the SQL operators. The quality of the statistics for the objects referenced in the statement greatly affects the quality of the plan. Statistics maintenance is a challenge that all DBAs must face in order to prevent the execution plan from becoming suboptimal. This presentation includes information on the features introduced in Oracle Database 11g to improve the quality and efficiency of statistics gathering as well as strategies for managing statistics in various database environments. These strategies include how to use dynamic sampling and when, and how to manually set statistics directly versus collecting them.

Oracle NoSQL Database is a commercial-grade, general-purpose NoSQL database using a key/value paradigm. It allows you to manage massive quantities of data, cope with changing data formats, and submit simple queries. Attend this session to learn when to use a NoSQL Database and how Oracle NoSQL Database can be used to support web applications, acquire sensor data, scale authentication services, or support online services and social media.