Hotsos Symposium Speakers – Neil Gunther and Peter Stalder
Neil Gunther is a returning Hotsos Symposium presenter. Peter Stalder is a new Hotsos Symposium presenter.
Biography
Neil J. Gunther, M.Sc., Ph.D., is an internationally known performance researcher and consultant who founded Performance Dynamics Company in 1994. He is the author of 3 books and many technical papers and presentations on performance analysis and capacity planning. Dr. Gunther received the prestigious A.A. Michelson Award from CMG in 2008, and was elected a senior member of both ACM and IEEE in 2009.
Mr. Gunther has presented at prior Hotsos Symposia (2007 and 2008) as well as CMG.
Peter Stalder has worked for DataMind as an Oracle Application Developer. Since 1996, he has worked as a consultant with Oracle database environments at Trivadis in Zürich. He has been active in numerous projects in operation and optimization of business-critical databases. Mr. Stalder's focus is on application performance management, predictive performance management and capacity management with Oracle databases.
Mr. Stalder has presented at euCMG, DOAG and ukCMG.
Presentation Title
How to Quantify Oracle Database Scalability, Parts I and II
Abstract
Most discussions about Oracle scalability revolve around configuration recipes and tuning tips. As useful as this information is, it lacks one thing: a means for assessing the claimed performance benefits. This is because such recipes and tips are merely qualitative information. In a court of law, they would be called "hearsay evidence", but in IT, we elevate them to "best practices". In this presentation, we show you how to do better than these "best practices".
The first requirement for making quantitative assessments is measurement: controlled measurement, where possible. However, even measurements are only part of the story. Without specifying expected performance benefits, how can you know if the measured performance is optimal performance? Quantifying scalability for single instance or RAC also requires a performance model---the rest of the story. We present a complete performance methodology for quantifying Oracle scalability including RAC.
The universal scalability law (USL) is a performance model that incorporates all the important scaling factors:
- Equal bang for the buck (linear parallelism)
- Cost of waiting for shared resources (loss of linearity)
- Diminishing returns at higher loads (throughput ceiling)
- Negative return on investment (retrograde throughput) into a single equation that can be easily embedded into an Excel spreadsheet. Chapter 10 of Craig Shallahammer's book is based on an earlier version of this model, invented by one of us (NJG).
Part I of our methodology covers throughput scalability. Oracle DBAs, however, are often more interested in response-time scalability. In Part II, we show how the USL is actually an expression for an underlying queueing model that enables us to apply it to response-time metrics. Our quantitative methodology is also a valuable tool for determining consolidation stategies and capacity planning.
Presentation Materials
Presentation materials are available online to attendees only.
Schedule
The speaker schedule is as follows: