0

Performance Days 2019


Trivadis is very pleased to organize in Zurich another outstanding seminar with top guest speakers. This year’s focus will be on Oracle Database performance and query optimization. Since performance is not simply a product one can buy but rather the results of an accurate planning and a correct implementation, the seminar will present valuable techniques not only for troubleshooting performance problems, but also to avoid them in first place.


The attendees will have free choice between talks scheduled in two parallel streams. Performance Days will provide a lot of valuable and practical information to diagnosing, resolving and avoid performance problems in applications involving Oracle Database.
The speakers will cover topics such as:

  • New indexing and partitioning features introduced in Oracle Database 18c and 19c
  • Why it is of paramount importance to keep your SQL statements simple
  • The role of the infrastructure in Oracle Database performance
  • Using resource management to deliver predictable performance
  • Leveraging execution plans for problem solving
  • In-depth look at how Oracle Database processes joins and star queries
  • Using Oracle's kernel debug, diagnostics and tracing infrastructure as well as ASH for troubleshooting
  • Why you probably do not need to read a 10053 trace file
  • How to gather object statistics
  • Tracing parallel executions to better understand how it works
  • What parent and child cursors as well as adaptive cursor sharing and statistics feedback are




    2 days



    Oracle is a complex coordinated multi-process system with layers of background processes potentially on the critical path of your application’s operations. When everything works well, you won’t notice much wasted time or surprising wait events. However, when some critical background process starts falling behind enough, all kinds of problems will start popping up. Due to impact amplification on many foreground sessions, the top-level symptoms showing up in AWR and OEM may mislead you to troubleshoot the wrong thing. When following the links of evidence all the way to the deepest process in the (wait) chain, you can construct the topology of a complex performance issue and address the root cause. In this session you’ll see multiple troubleshooting case studies using ASH wait chain analysis, where the typical tools would have left the customer fixing the wrong problem.

    Avoid "Horror Queries": Keep Your SQL Simple and Fast (Dani Schnider)

    SQL is a powerful and efficient query language when used correctly. But inefficient and inappropriate SQL statements can cause massive performance problems. Instead of solving such issues with additional indexes, optimizer hints, SQL profiles or SQL plan baselines, a much better and more sustainable solution is often to rewrite the SQL statements and to reduce the complexity of the queries. In this presentation, some real-life examples of terrible "horror queries" will be shown and rewritten into simpler, more elegant and faster SQL statements. The goal of the presentation is to show how to improve the performance of a query by factors with appropriate SQL statements and powerful Oracle SQL features.
    With todays’ machines it is usually necessary to run multiple applications or roles on a single machine. This brings up the question of how you can ensure that the performance of the applications can be set up to be reliable and predictable especially when combining OLTP and DW type activity. This paper explores the options available with presentation and live demos to show how virtualization, instance caging, resource manager and multitenant can be used to deliver the required performance.

    Deep Dive into the Implementation of Oracle's Join Methods: Past, Present and Future (Shasank Chavan)

    Join processing in relational databases involves combining the output of two row sources (e.g. tables) into one result row source.  Most SQL analytic workloads are dominated by time spent in join processing because of the nature of the operation being heavily compute and bandwidth intensive.  Oracle has developed numerous traditional join methods over the years (e.g. nested loop, sort merge, hash join, etc.).  Some of these methods have fallen out of favor over time by the optimizer (i.e. past), others are time-tested and commonly selected by the optimizer to this day (present), and then there's a new class of join algorithms which leverage the latest In-Memory Columnar data formats to deliver the fastest join evaluation known in industry (future).  This talk will dive into the technical implementation details of Oracle's plurality of join methods (including specialized ones such as vector transformation which optimizes aggregations over joins), give insight into why the optimizer may choose one method over the other, and what the SQL and storage engines are actually doing under the hood to implement the join.
    You can't blame the optimizer unless you are sure you feed it with relevant statistics. There is the Auto Job which suits lots of cases, and there are some exceptions: volatile tables, partitioned tables, bulk load, histograms and bind variables, etc. You cannot gather all possible statistics every day, and you cannot rely on stale ones. Let's define a few rules, guided by the recent features: 12c Statistics Advisor and 19c real-time statistics.
    As databases get ever larger in size and complexity and with the Oracle Partitioning Option becoming more accessible, especially within Oracle Cloud deployments, the Oracle Partitioning option has become more common and a more important feature in Oracle database deployments. This presentation will look at some of the important indexing strategies that can be applied in a Partitioned Database environment in order to effectively maximise both database/application performance and database management practices, dispelling some common myths along the way. With prudent partitioning practices, it’s possible to achieve performance levels that aren’t practical without partitioning being applied. It will also discuss some newer features introduced in 12c R2 and 18c that simplifies the migration to an effective partitioning strategy.
    This session will give you a “quick overview” of the variety of the latest (open source) tools that Tanel regularly uses when looking into Oracle performance problems. The presentation will have 3 parts: First, updates to tools like Oracle Session Snapper v5 and LatchProfX. Second, SQL Tuning helper scripts like an extended execution plan display and visualization tool. The last section will introduce a new Linux tool called Process Snapper that can sample any process activity at Linux OS level similarly to Snapper for Oracle.

    Oracle Database 18c and 19c New Indexing Related Features (Richard Foote)

    The latest Oracle 18c and 19c Database releases have included a number of new indexing related features that help to maximise database and related application performance. This session will cover in some detail at some of these new features and capabilities, how they work and how they can be utilised to successfully improve performance. Some of the features to be discussed include the Memoptized Rowstore, Scalable Sequences, Index/Partitioning Enhancements, JSON Search Index enhancements and one of the big 19c features, Automatic Indexing. With regard Automatic Indexing, not only do we look at how it all works, but also what useful things we can learn regarding manual index management.
    This session will examine a couple of execution plans that have appeared on the Oracle Developer Community database over the last year, explaining what information can be seen from the initial request for assistance, what further information might be needed, and what database considerations might come into play that are outside the immediate scope of the information available in a plan. There will be no Powerpoint slides in this presentation as the medium is not appropriate for displaying plans of a "real-world" level of complexity.
    The result of a parse operation is a parent cursor and a child cursor stored in the library cache. Obviously, the aim of storing them in a shared memory area is to allow their reutilization and thereby avoid hard parses. But when can they be reused? The aim of this presentation is not only to discuss what parent and child cursors are but also, and foremost, to examine in what situations it is possible to reuse them. In doing so, topics such as bind variable peeking, bind variable graduation, adaptive cursor sharing and statistics feedback will be covered as well.
    An important SQL basics: join methods. The approach is live demo showing Nested Loop, Hash, Sort, Merge, with gdb breakpoints on the Oracle executable in order to see how they are processed, in slow motion. When you understand how it works, you know exactly which join method is good for your query. We will also show some internals of one of the most advanced technique of the optimizer: 12c Adaptive Plans. Additional optimization for joins will be covered: table elimination, partition-wise join.

    Successful Star Schemas (Mark Ashdown)

     Star schemas and star queries have a strong track record in delivering valuable business information from relational data . The standardised structure of the schema is well-supported by business intelligence and reporting tools from Oracle and numerous other well-known vendors. It has also encouraged many optimisations in the algorithms used by Oracle Database, with new capabilities being introduced to take advantage of the latest developments in hardware. Join this session with Oracle's Real-World Performance Group to learn some simple rules that help to ensure success through consistent and efficient execution of star queries, together with a live demonstration of the trade-offs between different technologies.
    The aim of this presentation is to focus on the key role played by the infrastructure in a stable and performing database environment.
    Like a skyscraper, a database environment requires a strong foundation. In addition, to fully exploit all available resources, a great synergy among all hardware and software layers is of paramount importance. This presentation covers different architectures and configurations. For each of them, it highlights the expected performance benefits.
    This is an opportunity to extend your know-how in diagnosing parallel execution and learn how to analyze PX trace files. PX tracing can be used to get detailed information about operations performed by the coordinator and its parallel execution slaves. We'll look under the covers of parallel execution to better understand how it works. You’ll learn how PX processes communicate, answer questions like “Why are PX slaves not being allocated” or “Why does my statement not execute in parallel?”.
    The Oracle RDBMS has implemented three APIs (generic debug, generic trace, numeric events) in its low-level kernel debug, diagnostics and tracing infrastructure since version 11g. This session provides an overview of these three APIs, how to get more information about its tracing possibilities/syntax and how to use/set them. In the end some real life client issues that were troubleshot with help of these three APIs are discussed. You should be able to define and set your own custom actions/events for ORA errors and/or for specific C functions and/or for troubleshooting performance issues after this session.

    It's almost true to say that the only time you should read a 10053 trace file is when you don't need to. It's also probably a fairly accurate claim that the only good reason for reading a 10053 trace file is to prove that you didn't need to read the 10053 trace file. In all my years of consulting and trouble-shooting I think I've only used the 10053 trace file a couple of times to demonstrate the cause of a performance issue and suggest a work-around - but having spent a silly amount of time generating and reviewing trace files for increasingly complicated statements I found it easy to pick the bit of the trace file I needed to examine to confirm a hypothesis and explain the evidence to management (and Oracle Support). In this session I'll be giving you plenty of reasons why you shouldn't be reading a 10053 trace file - but we'll probably end up looking at some bits of one anyway.


    26.-27. September 2019
    The working day will be from 9am to 5pm; and on the evening of 26th there will be an opportunity to share your experiences and opinions with the speakers and other participants in a less formal atmosphere as we meet over drinks from 5pm to 7pm.


    The seminar will be conducted in English.


    Target audiences of this event are performance analysts, database administrators, application developers, as well as consultants who want to improve their skills in managing performance or developing database-backed applications involving Oracle Database.
    Participants are expected to have a working knowledge of Oracle Database.


    Join us either live in Zurich or attend the sessions online in our virtual classroom. All sessions will be available as live stream for you. Requirements: access to the internet and headset. You will receive your personal account prior to the event. The account is valid for one person.

    Special offer

    Conference Ticket: 1,950 CHF

    Price for the 2nd participant from your company - 1463 CHF
    Price for the 3rd participant from your company - 975 CHF
    Price for the 4th and every other participant from your company - 488 CHF

    Online Ticket: 1,650 CHF
    Price for the 2nd participant from your company - 1238 CHF
    Price for the 3rd participant from your company - 825 CHF
    Price for the 4th and every other participant from your company - 413 CHF


    The event takes place at the Holiday Inn Hotel Zurich-Messe, Switzerland.


    Live in the Holiday Inn Hotel, Zurich-Messe, Switzerland: CHF 1,950 / .2083 USD
    This price includes lunches and refreshments during the breaks.

    Live Stream: CHF 1,650 / .1763 USD
    The live stream will be recorded. This price includes access to the recordings exclusively for you.

    Hotel Special

    As a participant of the Performance Days, we can recommend the event hotel Holiday Inn, Zürich-Messe for your stay. When booking at the hotel with the code "PDays19" you will receive a special rate of CHF 200.00 / room / night incl. breakfast, excl. city tax.
    For booking please contact reservation@hi-zm.ch or +41 44 316 11 31.

    Our Quality Promise

    Contact to our service center

    DE+49 (0) 711 90363245

    CH+41 (0) 584 595795

    AT+43 (0) 133 2353160

    FR+41 (0) 584 595454

    EN+49 (0) 711 90363245

    DA+49 (0) 711 90363245


    Virtual Classroom
    26.09. - 27.09.2019
    EN1'650.00 CHF
    Zürich / CH
    26.09. - 27.09.2019
    EN1'950.00 CHF
    •    Standard
    •    Guaranteed to take place
    •    The course will definitely take place if you make a booking
    •    There are no free places left on this course. If you still book it, we will place you on the waiting list.

    • This course is held in another country. Due to the different tax conditions, you will have to book it in a separate booking process. We apologize for any inconvenience.


    back to trainings
    0
    Trivadis, Bernd Rössler, Head of Training

    Bernd Rössler

    • Solution Manager Trivadis Training
    • DE: +49 (0) 711 90363245
      CH: +41 (0) 584 595795
      AT: +43 (0) 133 2353160
      FR: +41 (0) 584 595454
      EN: +49 (0) 711 90363245
      DA: +49 (0) 711 90363245

      E-Mail: Training@trivadis.com

      “I'll be glad to advise you on topics such as individual coaching, workshops, project support, and online training courses.”

      Your Bernd Rössler

    Arrange a training consultation now

    Damit wir Sie auch in Zukunft telefonisch oder per E-Mail über die Trivadis Angebote und Produkte informieren können, benötigen wir Ihre Einwilligung. Ihre Einwilligung ist freiwillig und Sie können diese jederzeit ohne Angabe von Gründen formlos telefonisch oder schriftlich oder per E-Mail unter info@trivadis.com widerrufen. Ihre personenbezogenen Daten werden von uns an ein CallCenter weitergegeben, sodass wir Sie auch auf diesem Weg kontaktieren und über Trivadis Produkte und Angebote informieren können.

    trivadis training satisfaction guarantee

    Guarantee of quality

    Following completion by the participants, every single training course is assessed at the levels

    • suitability of the training room,
    • suitability of the workplace,
    • functionality of the technical equipment, and
    • satisfaction throughout the entire course / seminar,

    and recorded in our "TRIVALUATION" feedback system for quality assurance purposes. This enables us to consistently ensure the high quality and satisfaction of our valued customers. We really appreciate the overall rating of 9.3 from 10 points awarded by our many enthusiastic customers.


    trivadis training success guarantee

    Success guarantee

    Trivadis guarantees the success of your training. Having completed the course / seminar, do you have any questions about the practical aspects? Would you like to repeat any of the exercises in the lab environment?

    Our success guarantee allows you to repeat individual days, or even the entire training course, free of charge for up to 6 months after attending a training course. You bring with you the course materials from the previous training course.