Human expertise is needed even in the clouds! Welcome to the SQL Optimization Virtual Conference, 2020 edition!
Some of you may be old enough to remember that I organized a couple of virtual conferences back in 2010! The first event was called E2SN Virtual Conference: Systematic Oracle SQL Optimization in Real Life. The four speakers were Cary Millsap, Jonathan Lewis, Kerry Osborne and Tanel Poder. All sessions were technical, practical & awesome as usual and Kerry even played us some guitar! (He forgot to mute his mic)
I realized that almost 10 years have passed since the last time – we have got to do this again!
I hereby announce the next event in the Virtual Conference series: Systematic Oracle SQL Optimization in 2020.
This conference is about the same core topics, includes all the original speakers, but with present-day state-of-the-art content. And this time we have Mauro Pagano too!
I changed the structure of the event a bit, to make it more like a conference, with plenty of time for Q&A chat and speaker panels & general discussion. As an experiment, we’ll even open a slack channel for the duration of the event, for further Q&A, commentary, jokes and just hanging out with the speakers.
As announced here, we’ll do two runs of this conference due to people’s schedules being potentially uncertain right now (you can attend both versions of this conference for the price of one):
|11:00-11:15||Tanel||Welcome & Introduction|
|11:15-12:30||Cary||A Richer Understanding of Software Performance|
|13:00-14:30||Jonathan||Trouble-shooting with Execution Plans|
|14:30-15:00||All Speakers||Panel: What Has Changed in 10 Years and What Has Not|
A Slack channel will be open for conference chat between the two days.
|11:00-11:15||Kerry||How to Stay Relevant|
|11:15-12:30||Tanel||Scripts and Tools for Optimizing SQL Execution and Indexing|
|13:00-14:30||Mauro||Chase the Optimizer Every Step of the Way|
|14:30-15:00||All Speakers||Panel: What Will Change in the Next 10 Years|
Updated: The early bird price is the same as we had 10 years ago, so it’s an even better deal now in 2020! Early bird pricing is extended to 22. April 2020 at 11:59pm US Eastern time.
|Ticket type||Price per attendee||Sign up!|
Once the early bird offer ends, the GA price will be $575 per attendee.
Back to top
Cary Millsap is an entrepreneur, leader, teacher, software technology advisor, software designer and developer, writer, and Oracle software performance specialist. His technical work is quoted in many Oracle books, Wikipedia, blogs all over the world, and numerous conference presentations each month. He has presented at hundreds of public and private events around the world, and his blog is read by thousands of people each month. He is published in journals, including Communications of the ACM. Cary wrote the books The Method R Guide to Mastering Oracle Trace Data (2016) and Optimizing Oracle Performance (2003), for which he and co-author Jeff Holt were named Oracle Magazine’s 2004 Authors of the Year.
Jonathan Lewis is a well-known figure in the Oracle world with more than 32 years experience using the software. He has published three books about Oracle - the most recent being “Oracle Core” published by Apress Nov 2011 - and contributed to three others. He runs a couple of websites and contributes fairly regularly to newsgroups, forums, and User Group magazines and events around the world. Jonathan has been self-employed for most of his time in the IT industry. For the last 21 years he has specialised in short-term assignments, typically of a design, review, or trouble-shooting nature. He runs seminars about using Oracle all over the world and has visited more than 50 different countries to talk about, or trouble-shoot, Oracle systems.
Kerry Osborne is a VERY experienced Oracle guy (i.e., getting old). He currently wears the CEO hat at Gluent, a data virtualization software company. Kerry is interested in large databases, distributed databases, cloud-based databases, performance, and boomerangs.
Tanel Põder is a long-time computer performance geek, consultant and an entrepreneur. He is working mostly with enterprise data platforms like Oracle, Hadoop, Spark and the underlying operating systems. Tanel has built a number of performance troubleshooting tools for Linux and Oracle. He is a co-founder of Gluent, a company that builds data sharing solutions for enterprises.
Tanel holds two patents in data virtualization space and is a co-author of the Expert Oracle Exadata book. He is currently researching machine learning, so he could replace himself with an AI bot and retire in 2030.
|Mauro Pagano||Mauro Pagano is a database engineer with special interest in Database Performance and SQL Tuning. He spent the best part of the last decade focusing on Oracle problems, from database tuning to applications development. Mauro is an active member of the Oracle community, always willing to help or mentor other peers, and he enjoys giving back developing free code and presenting at Oracle User Groups conferences. He is the author and/or maintainer of some well-received free tools (SQLd360, TUNAs360, SQLdb360) and the previous maintainer of other legacy tools like SQLTXPLAIN and SQLHC.|
People have some funny ideas about performance that just aren’t true. For instance, have you ever had a vendor tell you that a certain upgrade will “make everything some number of times faster”? It’s almost never true, and a little bit of understanding makes it clear why a statement like that usually can’t be true. This session explains some fundamental ideas about performance, using two simple Excel spreadsheets, in a way that you’ll probably never forget.
This is a broad-appeal session for everyone: business users, system owners, project managers, architects, DBAs, developers, …anyone who has ever had a conversation or even wondered about performance.
This session will describe a new, more accurate way of understanding application performance that will make you less susceptible to being disappointed, and help you better understand how you should be measuring and testing your upgrades and migrations.
There is a huge amount of information available in an execution plans that can help you understand why a query is doing more work and taking more time than you expect - especially if you can take advantage of the dbms_monitor package to pull the plan and its execution statistics from memory after running the real query.
There are cases, though, where it is necessary to look closely at the execution figures to understand - or get the clues about - where the work is going and what is being done. Sometimes the numbers are reported in the wrong place, sometimes it’s hard to tell whether the reported times for some parts of the plan are the cause or the effect of the total time taken, sometimes the counters for reads or writes or buffer visits don’t seem to make sense and you need to crosscheck with the session’s wait events (v$session_event) or activity (v$sesstat).
In this presentation we start with a brief reminder of query blocks, transformations, the interpretation of basic execution plans and how the optimizer stitches multiple query blocks together in a plan and pick up a few special cases of patterns in plans, including updates, deletes, analytic functions and match_recognize(). Finally we go to a few live examples to highlight the importance of looking carefully at the predicate information and the execution statistics.
Kerry Osborne has been been working with Oracle since version 2. That’s a long time. Things have changed a lot in that time and they are changing faster now than ever before. Kerry will be talking about where he sees the industry going and what he thinks DBA’s will be doing in the future.
Tanel has built a lof of new scripts, tools and techniques since the last Virtual Conference in 2010. In this talk he will guide you through diagnosing, troubleshooting and fixing a number of problem SQL statements with his optimization method and the scripts to support it. We will look into easily detecting cardinality misestimates, locations where the biggest “cost jumps” happen in a plan and where most of the execution plan inefficiency and resource usage comes from. These techniques will be applicable both to long running queries and short frequently executed OLTP-style statements. The goal is to provide you an easy-to-use toolset for systematically approaching SQL performance issues, regardless of the application type. As a part of that, Tanel will also show you a brand new tool for semi-automated recommendation of new indexes, based on the existing workload analysis (this is not the same thing as the Oracle 19c Automatic Indexing feature that requires Exadata).
One of the great advantages of using a query optimizer is getting your SQL translated transparently into something better! The Oracle CBO is capable of transforming a SQL several times even for just a simple statement. This is probably why analyzing a CBO trace (event 10053) is painful and scary to many: who likes to read 100k lines of raw trace? This session focuses on the mechanics behind the CQBT framework and shows an analytical approach to digest any CBO trace file decomposing it into smaller pieces, making it easy to analyze.
Back to top