Sunday Workshops

z/OS Quick Start Toolkit

Instructor: Robert D. Andresen

When: Sunday 8:30 AM - 12:00 PM

Whether you’re new to z/OS or on an unfamiliar system you need to know where things are and how they are configured before you can be productive. Once you have access to a z/OS system there is a methodology you can follow to discover common z/OS objects, such as PARMLIB, PROCLIB, catalog structures and TSO/ISPF options. This workshop will describe commonly available z/OS tools, commands and utilities to allow you to quickly become effective on the z/OS system.

A brief history
Common z/OS assumptions
TSO/ISPF primer
Dataset names and types
Edit tools & tricks
ISRDDN capabilities
Catalog structure
Security
PARMLIB settings
Procedures
Jobs, TSO and Started tasks
Executables


Robert Andresen is a Technology Specialist with CA. He has been working in performance management and monitoring for several software companies since 1997 and has been supporting mainframe systems since 1976. He is a co-author of the IBM Redbook: Linux on IBM @server zSeries and S/390: System Management. He holds a degree in Mathematics from the Illinois Institute of Technology. At CA he supports their suite of DB2 for z/OS tools. His areas of expertise include z/OS, DB2, Oracle, CICS, UNIX/Linux, Windows, Websphere MQ and Networking.

New Methods and Metrics for Enterprise Storage Performance and Capacity

Instructor: John Baker

When: Sunday 8:30 AM - 12:00 PM

Enterprise Storage Performance and Capacity Management

This workshop will discuss the challenges involved in managing Enterprise Disk Subsystems in a mainframe environment. How many RMF metrics may actually mislead us into believing that things are running just fine; and how the old rules of thumb simply don't apply anymore.

Recent architectural changes have shifted the most likely bottlenecks from the host systems down to the Disk Subsystem components and why new levels of visibility and automation are required to avoid painful service level issues.

A new methodology and complementary new metrics will be introduced and discussed.

In part 2, we will discuss the capacity planning challenges in both Mainframe and Open environments. Modern Disk Subsystems are incredibly powerful and complex machines. But there are many layers of technology between the host interfaces and your data. Most SAN environments have very good views into the server and switch layers, but little to no visibility into the Disk Subsystems.

Whether the challenge is an existing Response Time issue, or a budget decision on the most cost effective upgrade options, we will discuss how the use of a proven mathematical modeling tool can take much of the guesswork out of these decisions. Ensuring that the right choices are made to balance the needs of performance vs cost.

16 years in IT industry at a large international bank. 11 years specializing in mainframe performance, with a focus on I/O performance. Designed, implemented and maintained many critical projects such as WLM Goal Mode and GDPS/Data Mirroring. Extensive experience with many performance analysis tools and techniques at the hardware, OS, and application levels. Hosted many prior sessions at Share, CMG, etc. As the Support Director for IntelliMagic, North America, John is responsible for education and technical support for all of IntelliMagic's prospective and existing clients.

Capacity Planning for Retirement

Instructor: Dr. Thomas E. Bell

When: Sunday 1:00 PM - 4:30 PM

As we CMG members age, many of us are approaching retirement. We need to take actions for ourselves and our families to protect us when our financial resiliency is reduced and our needs have evolved. We need to do the necessary planning so that our financial capacity will make retirement possible.

What should you do about health insurance and life insurance? How much money will you need, and how reasonable is it to get that much? What will be the effects of demographics (e.g., the baby boom, increased longevity) and probable government responses? If you sell your house, where should you live? Will taxes destroy you, and what reasonable plans should you make to cope with them? Should you just hand over your assets to that nice young financial planner and hope for the best?

When we're between 25 and 40, we need to make plans and begin saving. By the time we're about 40 or 45 we need to implement plans, especially because demographic shifts promise to put real burdens on people who haven't retired by 2015.

A real retiree (Dr. Thomas Bell) and a real Certified Financial Planner (Joe Delano) will describe the challenge and make specific suggestions.

Dr. Thomas E. Bell received his Ph.D. in Management from UCLA many years ago, and was a consultant for about 25 years before he retired. He is a recipient of the A.A. Michelson award (in 1975), has served CMG as its Treasurer and it Secretary, and has repeatedly been on the CMG Board of Directors. He retired at the end of 2005.

Joseph Delano is a Certified Financial Planner who has been involved in the investment industry for many years. He is the Broker for both Thomas Bell and CMG.

Visualizing SLA Compliance and Risk Management

Instructor: Dr. James Bouhana

When: Sunday 8:30 AM - 12:00 PM

Formal Service Level Agreements (SLA's) are increasingly used for ensuring Quality of Service. SLA's comprise one or more Service Level Objectives (SLO), with the objective of meeting a target value. This workshop presents both business and technical aspects of SLA's, including.
- Terminology, conventions, and variations used in formulating SLA's,
- SLA's as a component of overall QoS provisioning, Service Management, and Risk Management,
- ITIL and other standards related to SLA's.
- Deriving Objective Functions that yield future SLA values required to meet an SLA target,
- Deriving a Challenge Index that quantifies the risk of missing an SLA target,
- Using an SLA Projection Model for forecasting future SLO values for achieving the overall SLA target.
An SLA Dashboard is used to illustrate the technology and its visualization. Illustrative examples are presented for the SLA's of System Availability and End-to-End Response Time. A drill-down framework is used, whereby top-level results contain links to more detailed results. The overall objective of the workshop is to enable attendees to answer the two questions: (1) Is an SLA on track to meet its target? and (2) What can be done to ensure compliance to target?


James Bouhana has extensive experience in capacity planning and performance management in both academic and commercial environments. He has been a frequent contributor to CMG conferences for 30 years. Dr. Bouhana is currently involved in research on advanced methods for performance data mining and SLA tracking and prediction. He has been active in recent years in several R&D projects at the Fujitsu Australia Systems Engineering Research Centre in Melbourne, where the work reported in this workshop was done.

Working Through z/OS CPU Measurements

Instructor: Peter Enrico

When: Sunday 1:00 PM - 4:30 PM

With the advent of specialty engines such as zIIPs and zAAPs, and the new z10 processor, measuring CPU time and processor utilizaitons are not as simple or straightforward as they use to be. Now when analyzing CPU consumed at the address space level, workload level, or LPAR level we need to be concerned what work is running on what types of processors, the speed of the different processors, Hiper Dispatch, and a whole host of other factors.

During this z/OS performance workshop, Peter Enrico will provide a real world practical understanding and insights into this key z/OS performance topic. Even if you’ve been working with SMF measurements for a long while, you are sure to learn something new. At a high level the topics to be discussed include:

- Level set of zArchitecture specialty engines

- CEC level and LPAR level measurements

- Workload level processor measurements

- Address space level processor measurements

Extra Offer : Send SMF Data Prior To The Workshop!

Peter will process any SMF data received at least 1 week prior to workshop. Peter will make available, to those attendees, tables and graphs of their own data. On an informal basis, Peter will be available during CMG week to those wishing to discuss their measurements.

For data gathering instructions, email Peter at Peter.Enrico@EPStrategies.com.


Peter Enrico has strong and diverse experience with the IBM zArchitecture platforms, and a solid background in z/OS, Workload Manager, Parallel Sysplex, UNIX System Services, and WebSphere e-business performance. Peter also has extensive experience measuring, analyzing, and tuning z/OS systems, Sysplexes, and subsystems. Peter's abilities extend beyond just performance. He is considered a strong project manager and a highly qualified and effective communicator and seminar instructor.

Peter regularly teaches z/OS performance classes and as a member of Enterprise Performance Strategies, Inc., he works as a performance consultant for companies interested in optimizing the performance z/OS systems.

Details of Peter's seminar schedule, his services, his software, and for past papers and presentations, please visit http://www.epstrategies.com.

How High Will It Fly? Predicting Scalability

Instructor: Dr Neil J. Gunther

When: Sunday 8:30 AM - 12:00 PM

All systems, whether physical or computational, have operational limits. An aircraft cannot exceed its flight parameters or it may stall. Exceeding the carrying capacity of a bridge can lead to its collapse. Physical systems exhibit a predictable critical point beyond which they collapse. The same can hold for computer systems but more commonly performance just degrades. Highly scalable applications are desirable for reasons other than performance, e.g., lower development cost, availability, simpler maintenance, flexibility and heterogeneity. How can we characterize these effects and incorporate them into predictable scalability before an application gets deployed? In this tutorial, we will identify and quantify the following critical determinants of good scalability: 1. Equal bang for the buck: ideal parallelism or concurrency 2. Diminishing returns: due to contention delays in the system 3. Negative dividends: loss of capacity due to coherency delays which leads to the *Universal Law Of Computational Scalability*. The course notes will contain new material which updates my book ""Guerrilla Capacity Planning"" and each attendee will have the option of obtaining a copy of the book at a substantial discount.


Neil Gunther, M.Sc., Ph.D., is an internationally recognized IT consultant
who founded Performance Dynamics in 1994. He has held positions at San Jose
State University, JPL/NASA, Xerox PARC and Pyramid/Siemens Technology. His
training classes have been given at both corporate and academic
institutions including AOL, Boeing, FedEx, Motorola, Stanford University
and Sun Microsystems. Dr. Gunther has written several books and is well
known for his CMG presentations. Currently, he is working on Quantum
Information Technologies. He is a member of the AMS, APS, ACM, CMG, and
IEEE.

Visualization and Analysis of Performance Data using R

Instructor: Jim Holtman

When: Sunday 1:00 PM - 4:30 PM

As performance specialists we are presented with a large amount of data, from various sources, which we have to analyze to determine the characteristics of a system. This requires that we read in the data, convert it to a representation that is easy to process (e.g., spreadsheets, matrices, vectors, etc.), perform various calculations (e.g., average throughput, regression analysis, response time by transaction, etc.), and present the data in various formats (reports, graphs, web pages, etc.) so that others can understand what it means.

This workshop will introduce the student to the Open Source software called R. This is a statistical programming environment that allows the quick interactive development of a script and then the script can be run in a batch environment for repeated executions. This is very similar to the way that the UNIX “shell” is used; you can interactively try out commands and then store them in a script. The workshop will provide the basics of the language and will use some performance data (‘vmstat’ and ‘ps’ from UNIX/Windows) to illustrate the commands. Scripts will be provided so that the student can implement these operations on their systems.

Although I am not a rocket scientist, I have shot down a few ballistic missiles in my 42 years in software development. I was a coop student at White Sands Missile range for 5 years. I worked at Bell Labs for 23 years and for Convergys for 14 years.

Did my undergraduate (BSEE) at New Mexico State University and graduate work (MSEE/CS) at Berkeley (I was there when Regan was governor and he called out the National Guard on campus to put down the riots - remember People's Park?)

Career spans COBOL, PL/I and IMS on the mainframes, building a real-time operating system for the Safeguard Antiballistic Missile system, hacking the UNIX kernel, operational support systems for AT&T, and billing and customer care systems for the telecommunications industry. I developed the architecture review process that was used throughout Bells Labs (for which I was named a Bell Labs Fellow) and have reviewed over 200 projects in my career looking at their architecture and performance requirements. Currently I am working as a ""retired"" consultant developing simulation systems so that a client can test out a process before making the change in the real-world.

I have taught courses on systems architecture and ""working the numbers"" as they relate to performance on a project at both Bell Labs and Convergys.

Performance Management with Free and Bundled Tools - 2008

Instructor: Mario F. Jauvin & Adrian Cockcroft

When: Sunday 1:00 PM - 4:30 PM

This workshop includes new and revised content for 2008.

Abstract:

Computer system and Network performance data collection, analysis, modeling and capacity planning on any platform using bundled utilities and freely available tools such as Orca, BigBrother, OpenNMS, Nagios, Ganglia, SE Toolkit, R, Ethereal/Wireshark, Ntop, MySQL and PDQ.

Overview:

Capacity planning and performance management tools have been commercially available for many years. A new generation of freely available tools provides data collectors and analysis packages. As the underlying computer platforms and network devices have evolved, they have added improved data sources and have bundled free data collectors. Several open source and freeware projects have sprung up to collect and display cross-platform data, and with the advent of highly functional free statistics and modeling packages comprehensive analysis, modeling and archival storage can now be assembled. Free and bundled tools are of special interest to sites with very diverse mixes of systems, very large sites where licensing costs become prohibitive, and sites replacing a few large single systems with many more low cost horizontally scaled systems.


Adrian Cockcroft - formerly at Sun for many years, well known author and presenter. Now at Netflix.
Mario Jauvin - formerly at Nortel, now an independent consultant at MFJ Associates

Java Performance Analysis and Tuning

Instructor: Peter Johnson

When: Sunday 8:30 AM - 12:00 PM

Attendees at this workshop will benefit from my many years of doing Java performance tuning, both in our lab where we run industry standard benchmarks, and in our application excellence centers where I have helped tuning our customer’s real-world applications. I will cover the following topics:

* Understanding the garbage collector (GC), gathering and analyzing GC data, and using that to tune the Java heap.
* Understanding various GC algorithms, how to configure them, and how to decide which one works best for you.
* Working with the Java Management Extensions (JMX) to gather data about the performance of your application, including the use of VisualVM and other applications to monitor that data.
* Taking a holistic approach to Java Enterprise Edition (Java EE) tuning, where I show you what to tune in the OS, network, JVM, application server and application.

After attending this workshop you will have a concrete set of tools and ideas that you can apply to your Java application to improve its performance.


Peter started his computer career in August of 1980 working for Burroughs, programming mainframes in COBOL and Algol. He started working in Java in 1998, and was lead designer on projects such as a JDBC driver for the DMSII database that runs on Unisys mainframes. For the past several years he has been chief architect on a team that does performance analysis of Java applications on large-scale Intel-based machines (8 to 32 CPUs), and evaluates various open source software for enterprise readiness. In addition, Peter is a JBoss committer and is co-author of the book JBoss In Action, published by Manning. Peter speaks often on Java and open-source related topics at various industry conferences.

Best Practices of Performance/Diagnosis Tools in Windows Vista and Windows 7

Instructor: Dr. Kukjin Lee

When: Sunday 1:00 PM - 4:30 PM

Diagnosing performance problems and system bottlenecks is a complex and time-consuming task. A successful diagnosis requires in-depth system knowledge, a robust data collection mechanism, and comprehensive analysis tools. Windows Vista provides new innovations that will make diagnosis experience much more pleasant. The new Event Tracing for Windows (ETW) and Performance counter infrastructure provide a powerful instrumentation solution. In addition, the new Reliability and Performance Monitor (RPM) (formerly Perfmon) provides built-in diagnosis knowledge and richer UI controls that help users quickly gain insight into performance issues. Likewise, the newly improved Task Manager becomes a more valuable tool with added details and services integration. Windows Vista also introduces a new programming model called Performance Logs & Alerts (PLA) for logging various types of performance data including performance counters, events, API traces, registry, and WMI data.

Join us and see how easy it is to instrument and use tools to drill down performance tools. This workshop will focus on the best practices of diagnosis/monitoring solutions of the Windows Vista. In addition, we will demo the latest innovation in Windows 7 that leverages Windows Vista experience. The demo includes a new Powershell-based diagnosis solution and the new Resource Monitor.


Kukjin Lee is the Program Manager in Windows Core OS Fundamentals team at Microsoft. He has been responsible for performance/diagnosis tools in Windows, such as Task manager, Performance Monitor, and Resource Monitor. He also drove several infrastructure developments in Windows Vista, including the Performance Counter and Performance Logs and Alert (PLA). Before he joined Microsoft, he was an active researcher in the area of performance visualization. He received a Ph.D. in Computer Engineering from the Iowa State University.

Alex Bendetov - BIO -
Alex joined Windows Server Performance team at Microsoft in 2001 and has been working on performance related technologies ever since. Projects he worked on included analyzing and tuning performance of various Windows components and subsystems such as NamedPipes, Registry, FileServer and ActiveDirectory. Currently Alex is leading a team responsible for core event tracing infrastructure in Windows (ETW) and supporting libraries and tools (MessageCompiler, TDH, tracerpt).

Performance Management of SOA-based Enterprise Applications

Instructor: Dr. Odysseas I. Pentakalos

When: Sunday 1:00 PM - 4:30 PM

Service Oriented Architecture (SOA) has now emerged as the architectural style of choice for developing enterprise applications that will result in the sought-after benefits of reusability of software components, flexibility, and business agility. Enterprise applications are not developed but rather composed of individual, service components, each of which encompasses a unit of business functionality. A service, which may encapsulate legacy functionality or a recent requirement, may be deployed anywhere across the enterprise and possibly at multiple locations.
This session will present the challenges of managing the infrastructure that must be in place to support a scalable SOA implementation along with real-world solutions to some of these challenges. After an introduction to the characteristics and benefits of SOA-based enterprise, we will discuss the challenges that arise in managing the operation and performance of this environment which consists of services deployed across many heterogeneous platforms and where the dependencies between the services change constantly. We will then present approaches and solutions that an organization can utilize today to build the management layer that is needed in order to achieve a successful SOA implementation.

Odysseas Pentakalos is Chief Technology Officer of SYSNET International, Inc., where he focuses on providing his clients consulting services with performance management of computer systems and architecture of large distributed systems. He holds a Ph.D. in Computer Science from the University of Maryland. He has published dozens of papers in conference proceedings and journals, is a frequent speaker at industry conferences and is the co-author of the book Windows 2000 Performance Guide that is published by O'Reilly. Odysseas can be reached at odysseas@sysnetint.com.

Performance Tools and Problem-Solving in Solaris

Instructor: Bob Sneed

When: Sunday 8:30 AM - 12:00 PM

This workshop will provide a comprehensive overview of the tools available for performance measurement and management in Solaris OS environments. Discussion of the tools will be prefaced by some discussion of problem-solving strategies and some essential conceptual context for performance problem-solving in general.

Our tool survey will include the native tools provided in Solaris, as well as Sun's developer tools and 3rd-party tools. Some focus will be drawn to DTrace and the ecosystem of tools that has evolved around DTrace. We'll demonstrate how anyone can benefit from DTrace without being a programming guru. Drill-down coverage will focus on the most important basic tools in Solaris. Flashy inspiring demos will be given of some of the fancier tools for which drill-down exploration will be prohibited by time constraints.

Attendees will be provided with options for exploring the tools hands-on using their own notebook computers. A common lab system will also be configured for common use during the workshop.

The content is designed to empower practitioners and enlighten newcomers to the Solaris environment. It should prove valuable for attendees with diverse skills and interests.

NOTE: Companion to PM workshop More Performance Tools and Problem-Solving in Solaris""""

Dr. Darryl Gove works in the compiler team at Sun Microsystems on analysis and tuning of applications and benchmarks, and is the author of 'Solaris Application Programming' (Prentice-Hall, 2008). He blogs at http://blogs.sun.com/d/ .

Bob Sneed works in Sun's Performance & Applications Engineering group (PAE), and has authored several whitepapers and Sun Blueprints. He intends to blog more at http://blogs.sun.com/bobs .

Rick Weisner works on in Sun Solaris 10 Adoption and Web Infrastructure, with a focus on performance topics, Solaris internals, and C & C++ software development.

Solid State Storage and Your Database – Performance, Benchmarking, and Design

Instructor: Robert K. Williams

When: Sunday 1:00 PM - 4:30 PM

Solid state storage is quickly becoming main stream with uniform access times, lower power consumption, and higher performance. These features have a large impact on overall database response time and performance, as databases are traditionally more I/O bound than many other applications. This workshop walks you through the technology, products, and benchmarking in both OLTP and data warehousing roles. It will also look at issues such as capacity planning, and design considerations with solid state storage. It aims to focus on the concerns of those who run DB2, Oracle, and SQL Server so they have a strong idea on how solid state storage fits within their enterprise.

Rob Williams is a Certified Advanced DB2 Database Administrator and Application Developer. For the past 7 years, he has worked first- hand with government institutions and international large and medium sized corporations. He particularly focuses on database performance problems, design, migrations and implementations. Rob graduated from the University of Waterloo in Computer Science and Business with a strong emphasis on mathematics.

Hands on Workshop on Performance Prediction for Multi-tier Distributed Environment

Instructor: Dr. Boris Zibitsker

When: Sunday 8:30 AM - 12:00 PM

Objective

• During the workshop you will learn how to build and apply analytical models to predict the impact of workload and database size growth, new applications and hardware upgrade.
• You will use our Excel spreadsheet with exercises
• At the end of the workshop you will summarize results and prepare a report with capacity management recommendations.

Outline

1. Objectives
2. Description of exercises
3. Overview of the major factors affecting scalability, concurrency and parallelism of applications in a multi-tier environment with distributed Oracle RAC, DB2 and other DBMS
4. How to collect OS, DBMS and JVM performance measurement data
5. How to perform workload characterization 6. How to perform workload forecasting
7. How to build analytical queueing network models
8. How to predict the impact of workload growth, and evaluate the impact of hardware upgrades
9. How to set a realistic SLO and SLA and organize proactive SLM
10. How to predict the impact of database tuning
11. How to predict the impact of new applications
12. How to predict the impact of the server consolidation
13. How to organize proactive performance management
14. How to prepare and present a Strategic Performance Management Report
15. Summary


Dr. Boris Zibitsker is the Chairman and CTO of BEZ Systems. He is responsible for research and development of proactive performance modeling technology.
During his career Boris has working as Capacity Planner, Adjunct Associate Professor at DePaul University and vendor. Boris founded BEZ Systems in 1983. Boris and his colleagues have developed proactive performance management tools which are used by many corporations. Dr. Zibitsker has consulted with many of the largest enterprises in the world and taught seminars in North and South America, Europe, Africa and Asia.