SPRING CONFERENCE 2007

Hosted by Lockheed Martin

May 17, 2007

at Lockheed Martin, Building 157, Sunnyvale, CA

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

All conference attendees, vendors, speakers and NoCOUG staff must RSVP prior to May 10th in order to receive a Drive-In Pass at the Visitors Center. (click HERE for a map). The same forms of identification are required as last year.

US Citizens: State or Federal Government photo ID (i.e. Drivers License, U.S. Passport).

Permanent Residents: Green Card along with State or Federal Government photo ID (i.e. Drivers License).

Foreign Nationals other than Permanent Residents: Due to more stringent security requirements imposed by Lockheed Martin and the Department of Defense, Foreign Nationals other than Permanent Residents are not allowed entry to the Lockheed Martin facility and therefore can not attend the meeting this year.

Also note: Cameras of any kind are prohibited.  This includes, but is not limitted to, camera phones and PDAs with cameras.

Possibly an executive or even a co-worker in your company considers Oracle still just a database company. And if it is someone within your management chain, it might be impacting your role without your knowledge of it. Oracle technology continues to change as well as its use and management. Database growth, business and Services Oriented Architecture continue to alter the role of the Oracle Administrator. Attend this presentation and learn how Fusion Middleware is changing the role of the database administrator as well as other technology professionals. New technology presents new opportunity if you are prepared to take advantage of it.

Learn how to optimize both the performance and maintainability of your SQL statements in PL/SQL programs. One of the hallmarks of Oracle is the ease with which you can write SQL statements in PL/SQL. It is also, quite ironically, one of the most dangerous and problematic features of PL/SQL: it is simply too easy to write SQL in PL/SQL. This session offers some key best practices for when, where and how to write SQL statements in your PL/SQL programs. Beyond that, we will cover the most important PL/SQL-specific features for SQL optimization, including FORALL, BULK COLLECT, and table functions.

Building modular, reusable code doesn't happen by chance. We have to learn how to "just say no" to spaghetti code that is impossible to read, maintain, or enhance. In this presentation, Steven will show you how to take advantage of packages, local subprogram units, cursor variables, table functions, autonomous transactions, and more to craft small, reusable units of code that are easy to maintain and to test.

The advances in hardware and Oracle automation features are going to have a huge impact on the job duties of the Oracle professional. This presentation explores the two-day DBA training program, the rapid server consolidation movement and other industry trends to show how the job of the DBA will move beyond compartmentalized duties and into a broader spectrum. The Oracle Professional of the 21st century will be relieved of the tedium of monitoring and tuning and be free to concentrate on other important database administration activities. This fun and interesting presentation will give the attendees a look into how their jobs are going to change, sooner than they think.

With each new release of Oracle, the cost-based SQL optimizer (CBO) becomes more intelligent and Oracle10g is no exception. This presentation shows the major enhancements to the cost-based SQL optimizer in Oracle10g with a focus on the tuning tools that affect the behavior of the CBO. Topics will include the CBO Oracle parameters, the use of schema statistics by the CBO, and how the CBO evaluates complex queries and derives the optimal execution plan. We will also cover the use of the new Oracle 10g SQL profiles and discuss cursor sharing and automatic query re-write.

There are many ways to use Oracle wait events for performance tuning of an Oracle database, but often there is confusion on exactly what the data means. The issue typically centers around the fact that wait event data is analyzed at the wrong level or the collected wait event data is not detailed enough. This presentation will focus on these problems and review several real-life case studies of using wait event data coupled with Wait-Time based performance analysis to solve the most difficult performance related issues.

The AWR scripts that ship with Oracle 10g provide a lot of useful information, but there's a wealth of information contained in the AWR tables that isn't visible through these scripts. At our company we generate a number of additional reports as part of every performance test run against one of our applications and use that information to find both application and database problems. The reports that we get from the AWR tables include information such as a comprehensive list of the top SQL and query plans by various criteria, a list of all of the tables in our application and the query plans that reference each of the tables, and a list of all of the indexes in our application and the query plans that reference each of the indexes. This presentation will go over some of these reports, the SQL behind the reports and how we use the reports to tune our applications and improve our performance testing.

Oracle Streams is a powerful and flexible way to automatically replicate changes. Although simple in concept, the actual operation of Streams has lots of "moving parts." This means that troubleshooting replication issues can be very complicated, because there are lots of places to look when things don't go right. In this presentation Brian Keating and Chris Lawson tackle the bewildering task of monitoring and correcting Streams problems. They suggest various ways to monitor the Capture, Propagate, and Apply processes, and what to do when things go wrong.

Oracle Real Application Clusters (RAC) has been steadily gaining momentum in the market with new and old customers considering RAC implementations. Both DBAs and managers need to understand some basics about RAC and how it is managed. There have been many mistakes made by those that attempted to create RAC environments without having a solid understanding of the architecture. This session will focus on the technical architecture of the RAC feature with an emphasis on those areas that are frequently misunderstood.

Cary Millsap has sounded the alarm about killer "skew" that prevent us from discerning the performance problems through the fog of StatsPack and AWR data. In this presentation we will refine a method (based on LOGOFF triggers) due to Richmond Shee which helps us find killer skew and illustrate the method with a real-life example.

This presentation will introduce Oracle Database Administrators, Unix System Administrators and Storage Managers to Oracle storage options. The focus will be on introducing Oracle Automatic Storage Management and Oracle Clusterware to individuals new to Oracle storage options. As Oracle storage solutions continue to evolve, it is important for all IT individuals working with storage understand the pros and cons of different storage options. Attendees will learn the fundamentals of Oracle Automatic Storage Management and Oracle Clusterware from a storage perspective and not just an Oracle perspective. This is an excellent presentation for individuals new to Oracle storage solutions.

Any time data is exchanged between distinct and dissimilar organizations, a relationship exists with a variety of duties, expectations, and technical issues. This relationship must be managed. When one party seeks to acquire data from another, there are a host of issues which must be addressed unambiguously. The acquisition of data (particularly from business production systems) usually places some kind of burden upon the data originator/supplier, both for initial extract, and recurring updates. Issues of architectural differences, architectural stability, scope, data quality, replication techniques, permissible usage, liabilities -- all these must be addressed. Of particular concern are incremental extractions of data after the first big bulk transfer. It is crucial to understand the difference between updates and corrections. It is also important, if the incremental transfers occur over a longer period of time, to detect changes in scope, architecture, meaning, and quality of the incremental updates.