Summer Conference 2007

Hosted by Chevron

August 16, 2007

at Chevron, San Ramon, CA

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

Leng Tan will discuss the manageability features in Oracle Database 11g which are designed to help organizations easily manage enterprise grids and deliver on their users' service level expectations. Oracle Database 11g introduces more self-management and automation that will help businesses reduce their system management costs, while increasing performance, scalability, availability and security of their database applications. During this keynote presentation, Leng will detail many of the new manageability capabilities in Oracle Database 11g, such as automatic SQL and memory tuning, partitioning advisor for tables and indexes, and enhanced performance diagnostics for database clusters. Other highlights of this session include an overview of Oracle Database 11g's unique new 'change assurance' features designed to help businesses significantly reduce the time, cost and risk associated with system changes such as hardware or software upgrades, as well as configuration changes.

RDBMS technology and the RDBMS market have undergone huge changes in past years with more change to come. What are the key industry and technology trends influencing the future directions for relational databases and how do these trends affect RDBMS professionals? In particular, we'll consider the effect of grid technology, virtualization, changing software architectures, open source, outsourcing and security considerations.

This presentation provides an overview of Oracle Database10g Streams technology. The session offers an introduction to the basic concepts and architectures of Oracle Streams. It discusses the following implementation topics: Basic Objects, Rules, Capture, Instantiation, Propagation, Apply, and administrating the Oracle Streams environment. This session provides a step-by-step setup of an Oracle Streams environment. It also compares Streams technology with Data Guard and Advanced Replication.

Oracle Database 10g has introduced a number of "automatic" features such as ADDM (Automatic Database Diagnostic Monitor), AWR (Automatic Workload Repository), ASH (Active Session History) and Server Generated Alerts (SGA) that are based on Server Metrics. These features are actually based on internal infrastructure components consisting of V$ and X$ views that existed even in Oracle 8i. Once we have understood how Oracle Database 10g uses such components in these "Automatic" features, it is a simple matter to "back port" much of this functionality to the older versions. In this presentation, we will see some practical examples of this back porting and point to other areas that can benefit from this. We will also list what features cannot be "back ported" and why.

In Oracle RAC parallel processing environments supporting applications with different performance, resource utilization and data usage profiles, an increase in the number of users and volume of data, and implementation of new applications increase the contention for resources and affect performance of applications differently.

Oracle RAC has several self-healing features helping with performance tuning, but how do you justify workload management and operational decisions, RAC hardware upgrades, server consolidation decisions, and changes in software parameters to control the level of concurrency and parallelism? Interdependencies between application servers and DBMS servers in a multi-tier environment make life even more complex. It is difficult to avoid surprises if you do not know what to expect.

In this presentation we will focus on the role of workload characterization and performance prediction in justifying strategic, tactical and operational decisions, and organizing a continuous process of proactive performance management during RAC application life cycle. We will review how to set realistic SLO and SLA, justify management decisions, verify if actual results are close to expected, and if not, how to make proactive corrective actions.

Oracle performance tuning has matured over the past decade with ad-hoc and inaccurate tuning approaches giving way to more empirical and effective profiling approaches. However, there are still many circumstances in which performance tuning addresses symptoms rather than the underlying causes of poor performance. In this presentation, a systematic approach is presented which focuses on systematically optimizing the Oracle RDBMS to avoid inefficient or needlessly costly tuning outcomes.

2007 marks 30 years of Oracle helping businesses and governments manage their most valuable asset -- information. It's the lifeblood of all your applications; from traditional transactional systems and large data warehouses to Web2.0, XML, and other applications. During this presentation, Penny Avril will discuss today's IT challenges and how Oracle Database 11g can help manage user's service-level objectives while reducing the cost of computing.

Average Active Sessions, the golden mean of performance metrics, is a simple, powerful and elegant way of displaying the multidimensional data found in Active Session History (ASH). Between the dense approximate samples of ASH and the broad accurate snapshots of Statspack, we find a simple metric that can easily determine database health. In this presentation find out how to determine Average Active Sessions from Statspack, AWR or ASH and find out how to use the metric with Enterprise Manager or by hand by mining ASH data.

RMAN in the Trenches is a sharing of discoveries, as a result of implementing RMAN over a period of several years. It is not a high level overview of features, and it is not a "how-to" cookbook. Intermediate topics are discussed, along with some basics. Beginners can benefit by having a better idea of what to watch out for. In Part II, topics include corruption detection, metadata, and flashback. We'll also have a sampling of the The Good, The Bad, and The Ugly.

The SQL trace facility helps the DBA diagnose performance problems and make applications run faster. Tracing works well in client/server environments or two-tier architectures where there is a one-to-one correspondence between a database session and an end user session. In such systems, you can activate SQL trace for one user's Oracle server process and get a detailed accounting of all statements executed in that user's session. Unfortunately, the SQL trace facility never worked well in connection-pooled environments or architectures that use an application server to multiplex many end user sessions into a smaller number of database sessions. Beginning with Oracle 10g we have the ability to trace an end-user's session, even though connection pooling may cause each request by that user to be served by a different Oracle server process. In this presentation we will walk through a real-life example of tracing an individual end-user's actions as they use a web-based application. Although each request by the end user might be handled by a different Oracle server process, and each Oracle server process will also handle requests from other end-users, all requests by our target end-user get traced and reported cohesively in one TKPROF report. The technique is not hard to use, once you understand it. By the time we walk through the entire process and look at real output in this presentation, you will know how to trace individual end users in your connection-pooled environments.

To successfully plan and execute an SOA project requires a methodology with SOA specific techniques to complement your standard software delivery practices. This session highlights the unique aspects of a methodology for implementing SOA projects, including service portfolio planning, service design, and business process analysis. Learn how to maximize the benefits from your SOA projects with these techniques!