Hotsos Symposium Speaker – Luca Canali

Luca Canali is a new Hotsos Symposium presenter.

Biography

Luca CanaliLuca Canali is a Senior DBA for CERN in Geneva, with 14 years of experience using Oracle products in the sectors of research, finance and telecom. Since 2005, he has been working at the start-up and operations of the database services for physics at CERN. Luca has presented at Oracle User Groups conferences in the recent past on the topics of ASM internals, RAC, DB storage performance, Oracle compression, ACFS and Data life cycle management, Active Data Guard.

Presentation Title

A Latency Picture is Worth a Thousand Storage Metrics

Abstract

The production database is suffering from degraded performance due to poor I/O response times, while storage monitoring shows that all is OK. A new storage system claims hundreds of thousands of IOPS, but fails to perform for our specific needs. Troubleshooting these types of issues can prove to be quite challenging. In many cases, storage performance metrics are readily available and aplenty but at the same time do not address the root causes of the problems. Storage latency data, on the other hand, is often unavailable or overlooked, while it is one of the most useful sources that we can exploit to understand storage behavior.

This presentation will focus on storage latency data: collection, visualization, analysis and its applications to understanding and tuning databases and their storage. We will explore the latency data sources available in the Oracle ecosystem: from Oracle's wait event histograms to DTrace and more. We will see how ad-hoc visualization techniques and heat maps play an important role in making sense of latency data. Finally, we will illustrate how a set of free tools the author has developed for collecting and visualizing latency data make this type of analysis immediately available to the interested Oracle practitioners.

Presentation Materials

Presentation materials are available online to attendees only.

Schedule

The speaker schedule is as follows: