- An EPS Seminar -
Additional Information:
Click here to visit our seminar schedule page for currently scheduled class dates and locatio
ns.
Click here for registration and pricing information.
Click here for travel information.
Hardcopy of Abstract and Outline:
Seminar abstract and Outline
Seminar Overview:
When was the last time you rolled up your sleeves and revisited and re-evaluated your WLM environment and goals? This 4-1/2-day seminar is for any performance analyst, capacity planner, or systems programmer responsible for the MVS Workload Manager and for revisiting and reevaluation of goals on a regular basis.
Please do not hesitate to contact Peter Enrico if you have any questions regarding this seminar or others.
A Week of Reviewing Your Own Data!
Prior to class we instruct you on what data to bring to class from your own systems. Then, while in class, you will also be spending a week doing an actual WLM performance Health Check® using the data you supplied. So if you have not done a WLM goal mode Health Check® in a while, or have some performance nagging questions, this is a perfect opportunity for combined instruction, time doing analysis, and an opportunity to get input and recommendations from a leading WLM performance expert (and past WLM developer) – Peter Enrico.
The following information can be found on this page:
Dates and Locations
Instructor
Seminar Abstract
Seminar Participation
This seminar is for you...
Prerequisites
Seminar Highlights
Seminar Outline
Seminar Instructor:
Peter Enrico [
Return to Top ]
Seminar Abstract:
This 4.5-day seminar has the most complete and comprehensive materials on the subject of Workload Manager.
Even if you've been in goal mode for a while, this seminar is sure to give you a deeper understanding of WLM functions and abilities. You'll learn considerations for when to re-visit and re-evaluate goals over time. You will also learn how to better monitor, tune, and control how WLM manages your workloads. Since WLM is more than just understanding service classes and goals, you will also learn about the many other WLM functions that your system could take advantage of for workload balancing and systems management.
Peter is a recognized expert on WLM, having worked in the MVS Workload Management algorithm design and development teams during his tenure at IBM, and having helped many clients convert to goal mode. He has also helped many more to tune their service definitions and exploit new WLM functions.
And remember... as with all our performance seminars, during class you will be analyzing your own data during the class exercises. Peter will also provide you real-time feedback. So please remember to think of this class not just as a week of instruction, but also as a week of doing a real performance analysis.
This seminar is for you…
- if you are the one responsible for the z/OS Workload Manager (WLM) at your shop.
- if you are responsible for performance measurement, management, and evaluation of your z/OS and OS/390 workloads and systems.
- if you want to learn more about WLM, how it works, and how it manages the different workloads on the z/OS systems.
- if you want to learm about basic performance evaluation, or how the z/OS system manages workloads.
[
Return to Top ]
Prerequisites:
A basic understanding of z/OS (or OS/390) and RMF are assumed.
[ Return to Top ]
Seminar Highlights:
This seminar begins with a review of the basic concepts of z/OS Workload Manager (WLM). Do not think of this section as just a rehash of an explanation of the very basics. This section will provide you with detailed information and understanding of the WLM constructs that are not commonly published or discussed.
Understanding how WLM works and manages workloads is a big first step towards evaluating your WLM service definition and goals. So after the basics are covered, the seminar then discusses how the z/OS Workload Manager actually works. This section provides some insight into the WLM algorithms and methods used by WLM to solve delays to improve response times and velocities. This section will also provide insight into the often-misunderstood methods that WLM uses to manage various workloads such as CICS, IMS, WebSphere, DB2, and more. This section also includes a discussion on enclaves and application environment and how these constructs are used by WLM to manage distributed workloads.
Next the attendee will learn how to effectively use the WLM controls of service classes and goals. This section is full of many recommendations targeted towards specific subsystems and workloads. This lesson also instructs how to monitor workloads being managed by WLM.
Migrating to goal mode was just the first step. Over time, however, your workloads change or are added; your systems change; your hardware changes. Such changes should cause you to reevaluate your WLM goals. Maybe you will need to add or delete a service class here or there. Maybe a new velocity goal is in order. This section of the seminar will help you to start thinking about how and when to revisit goals over time.
The seminar closes out with a number of miscellaneous topics relating to how workload manager makes possible workload balancing in a multi-system Sysplex environment.
What this seminar does not cover is the basics of using the WLM application. The WLM application is an ISPF front end that is used to define the WLM service definition. This seminar is not a lesson on how to maneuver through the ISPF panels. A demo could be provided if requested.
Performance measurements, analysis, and tuning considerations are discussed throughout. A key strength of this seminar is that Peter does not just teach tuning recommendations, but why each recommendation exists. Seminar materials have extensive notes for later reference.
[ Return to Top ]
Seminar Participation:
Prior to the seminar, enrolled students will be sent a note suggesting what configuration data and measurement reports should be collected and brought to class. The purpose of this information is two fold. First, it ensures that each attendee is familiar with their current or planned Workload Manager environment. Second, reports brought to class could be used to reinforce concepts learned. Peter is available before class, during breaks, and after class to discuss with each student their own environment.
[ Return to Top ]
Seminar Outline:
The following is a high level outline for this seminar. A detailed seminar outline is available upon request.
Part I: Introduction and Review of the WLM Concepts and Controls
Don't think of this overview of WLM concepts and controls as being the same material you've been hearing of from IBM, and others, for the past number of years. This overview is designed not just to level set with the basics, but also to provide you with some additional information you may not already know.
- Service Definitions, Service Policies, Workloads, Service Classes, and Service Class Periods
- Goals - Response Time goals, Discretionary goals, Velocity goals, System goals
- Report classes - heterogeneous and homogeneous
- Classification rules
- Report classes
- Resource Groups
- Special protections for critical workloads
- Miscellaneous topics
Part II: Understanding How Workload Manager Works
A big first step towards evaluating your WLM service definition and goals is gaining an understanding of how WLM works. This section of the seminar will help you with this understanding to allow you to better evaluate and tune your WLM service definition and goals.
- Workload Manager algorithm basics, or - 'How WLM works'
- Basic terminology
- Performance Index
- Delays and Using indicators
- Receivers, Donors, and Resources
- Receiver value, Net Value
- Plots and histories
- Policy adjustment flow control
- Resource adjustment flow of control
- Introduction to resources that WLM can manage
- Introduction to WLM algorithms
- The basics of WLM algorithms to solve MPL delays, CPU delays, Storage delays, I/O delays, etc..
- WLM Management of address spaces, enclaves, and application environments
- Understanding address space management
- WLM Management of distributed workloads (generic discussion)
- Understanding enclaves and application environments
- WLM Management of CICS and IMS workloads
- Understanding WLM CICS and IMS server management
- Managing based on transaction goals versus region goals
- WLM Management of DB2 workloads
- Includes DB2, DDF, Store Procedures, DB2 Query Parallelism
- WLM Management of web and e-business workloads
- Includes WebSphere for z/OS and IBM HTTP Server
- Multi-tiered WebSphere applications
- Web centric WebSphere applications
- EJB centric WebSphere applications
- WLM Management of WebSphere applications connecting into back end systems such as CICS and DB2
- The flow and control of a WebSphere transaction on z/OS
- UNIX System Services
- WLM Management of Batch Workloads
- Includes a discussion on WLM batch initiator management
Part III: Effective Use of Workload Manager Controls
- Understanding Workload Manager measurements
- How to answer the basic performance monitoring questions when monitoring your workloads.
- Understanding Address space oriented measurements
- Understanding Enclave oriented measurements
- Understanding CICS and IMS transaction measurements
- Effective Use of Workload Manager controls
- Evaluating goals for started tasks and system address spaces
- Evaluating goals for traditional workloads
- Evaluating goals for on-line transaction processing workloads
- Evaluating goals for distributed and e-business workloads
- Evaluating goals for new / non-typical workloads
- Setting up report classes
Part IV: Revisiting Workload Manager Goals Over Time
This section discusses how to revisit goals over time. The techniques discussed in this section are continuously evolving and being developed. When possible, specific techniques are taught. Otherwise, high-level general concepts are discussed.
- Introduction to revisiting goals over time
- When to revisit and re-evaluate goals (and 'how' when possible)
- Improperly set goals
- Goals regularly being missed
- Planning for environmental changes
- Exploitation of new WLM functions
- Adapting to changes to WLM and the system
- Improperly tuned system
- Business priorities and objectives change
- When exploitation of non-WLM functions influence performance
Part V: Miscellaneous Topics and Technology Update
As time permits, this section of the seminar covers a range of miscellaneous topics.
- Latest announcement and technology update
- Scheduling Environments
- Batch Initiator Management
- Intelligent Resource Director (IRD)
- I/O Priority Management
- Latest storage management agorithms
- Workload License Charges (WLC) considerations
- Latest changes to resource groups
- Miscellaneous topics
[
Return to Top ]
Additional Information:
Click here to visit our seminar schedule page for currently scheduled class dates and locatio
ns.
Click here for registration and pricing information.
Click here for travel information.
[
Return to Top ]
