Central Ohio DB2 User Group
(CODUG)
September 16, 1998 Agenda

Meeting at Nationwide Plaza 1 Heritage Room (Map)

Amdahl's Global Information Sharing product - by Ruby Stepansky

With an increase in managing numerous data stores, of which may exist in different databases and different systems, IT is faced with critical business challenges: How can this information be shared effectively? How does this information stay current, consistent, and synchronized? How does the enterprise manage this information?

Data replication, copy management, and data extracts resolve some of these issues. However, to effectively manage all of these operations and ensure information consistency, an enterprise needs a unified information sharing model with a single point of control that manages and monitors processes.

The Global Information Sharing(GIS) model will be discussed as a foundation for unifying data mapping, distribution routing, and synchronization scheduling for all data movement operations in the enterprise. The model, in essence, provides the glue between diverse platforms in the organization. While data replication is the fundamental technology supporting the model.

Ruby Stepansky is a Senior Technical Consultant for Amdahl Corporation and has presented at local, national and international RDBMS user groups such as GUIDE and IDUG in areas of performance, heterogeneous data replication, and application and system tuning. He specializes in client/server design, logical and physical database design, replication strategies and topologies, and data sharing and warehousing implementations.

DB2 Buffer Pool Tuning: Top Down or Bottom Up? - by Joel Goldstein

Programmers, Database Administrators and Systems Programmers need to understand the impact of DB2 Buffer Pool performance upon their applications. In this in-depth presentation, Joel will discuss the goals and objectives of DB2 Buffer Pool tuning and illustrate performance data that should wave a "red flag" indicating performance problems.

Outline:

• DB2 Data Sources, variables and indicators
  1. Systems statistics and Application Accounting Data
  2. DB2 performance trace data
  3. Variables and inter-relationships
  4. Your problems, just how bad are they?
• Top down tuning approach
  1. Tuning from available data sources
  2. How much can you tune - the pay back potential
  3. How long will it take?
• Bottom up approach
  1. Data interpretation
  2. Determining where the benefits will be
• Non DB2 data sources
  1. SMF/RMF data
  2. Understanding, interpreting, and using the data
• Summary, guidelines, and recommendations.

Joel Goldstein is an internationally acknowledged performance expert, consultant, and instructor. He has a dozen years of experience addressing DB2 design, performance, and capacity planning issues, and has assisted many large national and international clients with their systems and applications. He is a frequent speaker at DB2 user groups and has published more than two dozen articles on DB2 performance issues. Joel is currently a member of the IDUG Board of Directors. Joel is president of Responsive Systems, a software and consulting firm specializing in strategic planning, capacity planning, and performance tuning of online database systems.

Dynamic Statement Caching for UDB - by Phyllis Groezinger

Dynamic Statement Caching is a new function within IBM's Universal Database (UDB) for OS/390 (DB2/MVS) that is intended to speed up processing for dynamic SQL by saving the access plan for subsequent re-use. This function will be particularly applicable for warehouse and ODBC access, but will also help any application which includes dynamic SQL statements.

Understanding this function allows application developers, DBA's and DB2 systems programmers to construct higher performance within their mainframe and Client/Server architecture.

Phyllis Groezinger is Advocate and Senior DB2 specialist with more that 20 years experience with IBM database products. She is an expert with database design and construction, database recovery, problem determination, performance tuning and problem identification. She is also a network design and performance tuning specialist experienced with Client/Server architectures. Phyllis has consulted with IBM customers through out North America.