|
CMG2006 Published Papers
Please note: Downloads are available to CMG members only. Click here to learn more
Business Performance Management |
Fundamentals / Core Competency
Hot Topics |
Network / Internet |
Storage |
Unix / Linux |
Windows |
zSeries
Business Performance Management
Measuring and Projecting Power for High Density Computing
Dr. Thomas E. Bell
Power consumption of computers has been a minor concern for years, but even that level of concern has declined over the last decade due to increased efficiency. Lately, however, power demands have increased as the use of blade computing has become popular. Inexpensive measurement tools can determine the actual situation, and some vendor specifications can be used to project demand. However, typical power engineers are not familiar with the changing characteristics of computers, so computer technologists need to be deeply involved in plans and computations in order to state requirements.
Download paper (Available to CMG members only, Click here to learn more.)
Achieving Business Agility with SOA: Governance & SLA Management of Shared Service Ecosystems
Dr. Jeffrey P. Buzen
Annie W. Shum, PhD
While the SOA promise of business agility leading to competitive advantage is truly compelling, there is a growing concern among skeptics and proponents alike that pursuit of this agile nirvana may turn out to be a fruitless exercise. This session takes an objective look at the key ingredients required to unlock the promise of business agility with SOA. The conclusion is sobering but will have far-reaching implications for tomorrow’s IT: namely, leading edge technology alone is inadequate. Service life cycle governance and SLA Management make the difference between success and failure of SOA.
Download paper (Available to CMG members only, Click here to learn more.)
Performance Reporting in the 21st Century - Changes in Scope and Direction
Gregory V. Caliri
Performance and capacity reporting have been critical enterprise functions since the start of commercial computing. The reporting role has changed through the years. In the 21st century, there are often three audiences -- the capacity, performance and business specialists. To further add to one’s dilemma, there are more tools and report platforms today. Skill sets of the reporters have changed. The specialist must learn to adapt presentations, build confidence or ''zen'' in each audience, and still manage to communicate effectively. A modular approach to reporting is invaluable.
Download paper (Available to CMG members only, Click here to learn more.)
The ABCs (or should I say, CASs) of I/T Chargeback
Robert E. Chaney
I/T chargeback systems have been in place for decades in the mainframe environment and in many companies they are deployed for the entire I/T department. There are also companies that have no chargeback system. This paper uses guidelines from the Cost Accounting Standards Board (CASB) as a means to understand the components of an I/T Chargeback System (ITCS), and explains how those Components interact to accomplish I/T chargeback.
Download paper (Available to CMG members only, Click here to learn more.)
Adding Value to Performance Management with Business Metrics
Scott A. Chapman
A great deal of system and human effort is expended to collect detailed technical measurements. These measurements are usually stored in some form of a Performance Database, or PDB. However, these data do little to help us understand the purpose of those systems: supporting our business. These basic performance measures, tell us very little about the systems’ business value. To understand that business value we need to collect data about the work being performed in business terms. That data could be stored, along with selected technical measurements, in a Business Metrics Database, or BMDB.
Download paper (Available to CMG members only, Click here to learn more.)
An Implementation of a Business Metrics Database
Scott A. Chapman
This session will present one possible BMDB design and discuss a simple web-based reporting architecture based on XML and Javascript that could be extended to other applications as well. Potential sources for both the technical measures and business metrics of mainframe applications will be discussed.
Download paper (Available to CMG members only, Click here to learn more.)
ITIL Capacity Management: More Than Charts Over Coffee
Rich Fronheiser
Many organizations are embracing ITIL as a model for best practices. This paper provides a brief background of ITIL and the ITSM processes, with particular attention paid to Capacity Management. The author also considers two different organizations and discusses Capacity Management as practiced in those environments, comparing and contrasting those practices with ITIL Capacity Management. Finally, recommendations are made for implementing ITIL Capacity Management in any environment, with additional focus given to the interfaces between ITIL Capacity Management and the other ITSM processes.
Download paper (Available to CMG members only, Click here to learn more.)
Monitoring, Availability, and . . . Maslow?!
Chris Greco
Abraham Maslow was a pioneer in the area of enlightened management and psychology. One of his contributions to the work of social scientists everywhere was his Hierarchy of Needs (HON). This familiar pyramid is used and applied everywhere from the boardroom to the backroom. Managers have relied on it for motivating their employees since the concept was founded. This paper will associate the HON to both the computer and the technician, as well as show that the processes that currently exist in our fields are also governed by this very pervasive and persuasive concept.
Download paper (Available to CMG members only, Click here to learn more.)
A Methodology For Determining Response Time Baselines
Charles Hoover
For a long time the 8-second rule has been the norm for setting response time on web pages. But how accurate is this rule in our new high-speed, broadband era of the Internet? By looking at the previous research done on user expectations and collecting response time data from a variety of data sources, it has been possible for us to come up with basic baselines. Then utilizing the Application Performance Index (Apdex) we were able to compare the response times of various pages to see how well they performed.
Download paper (Available to CMG members only, Click here to learn more.)
The Minimum Daily Adult - The Right Metrics & the Wrong Metrics
Denise P. Kalm
In capacity planning and performance analysis, we are inundated with metrics that purport to measure performance, but how we display them and understand them is what matters. Many metrics we take for granted are actually not that useful, and yet, permeate our world. Understand why the same-old, same-old metrics aren’t good enough, and what works better in this brief paper.
Download paper (Available to CMG members only, Click here to learn more.)
Transaction Processing using J2EE Application: Performance with Tens of Millions of Users
Mark M. Maccabee
We examine performance issues that came in the process of benchmarking, measuring and analyzing an application executing under WebSphere. A lottery application (tens of millions of users) was stressed using a custom made benchmark. We describe the development of the benchmark (that involved the use of JCA). We show the dynamic of J2EE standards and how much of it is in production. Application structure, deployment and performance are presented as we found them on real running systems.
Download paper (Available to CMG members only, Click here to learn more.)
Accountability for System Performance
Cary V. Millsap
Businesses are commonly organized in such a manner that the departments held accountable for making application systems fast and efficient don't have enough leverage to get the job done. The result is painful. Users suffer from slow applications, and system costs spiral while organizations expend energy deflecting blame. This session describes how the right measurements and a few process changes can produce a culture of performance accountability across departments, resulting in faster, cheaper, more efficient systems.
Download paper (Available to CMG members only, Click here to learn more.)
What Performance and Capacity Management People Need to Know About Finance
Chris Molloy
Performance and capacity management (PCM) personnel need to be aware of the financial implications of their actions. IT Business models such as ITIL have a financial management discipline included in the model which interacts with the ITIL capacity management discipline. The purpose of this paper is to describe the relationship between finance and capacity management, using the ITIL as a framework for this description. The paper will extend these relationships using projects that a capacity planner may lead. It will conclude with financial mistakes made in PCM, and how one can avoid them.
Download paper (Available to CMG members only, Click here to learn more.)
The Future of Peformance Management and Capacity Planning
Chris Molloy
Providing performance and capacity management (PCM) service is an integration of people, process, and products. The purpose of this session is to explore each of these areas with respect to future industry direction. We will discuss the labor strategy of PCM, a combination of on site, off site, near shore and off shore resources. We will discuss IT process models and their implications on PCM. We will discuss how the new product technology is affecting the PCM space. The session will conclude with recommendations on how PCM personnel should proceed, in response to these changes.
Download paper (Available to CMG members only, Click here to learn more.)
Virtualization - Inhibitors to Server and Storage Virtualization, and How to Mitigate Them
Chris Molloy
Virtualization is becoming one of the leading technologies being implement in IT environments in order to decrease costs thru increased equipment utilization. Virtualization can occur at the server, storage, network, and application levels. The purpose of this paper is to review the benefits of server and storage virtualization and some of the implementation technologies that are out in the industry today. The paper will then go into further detail on the inhibitors to implementing those virtualization techniques. The paper will conclude with mitigation recommendations for those inhibitors.
Download paper (Available to CMG members only, Click here to learn more.)
Application of Supply Chain Mechanisms to an On Demand Operating Environment
Sam Nokes
Dave Cohen
With the introduction of the On Demand Operating Environment, traditional capacity planning methodologies are being replaced by just in time resource allocation. This session examines the application of supply chain mechanisms to an ODOE. After reaching the ideal condition with a complete adaptation of the capacity offer to the dynamic capacity demand, the tasks of managing the ODOE systems are reduced to a simple order scheduling and to the logistical configuration and supervision (or monitoring) of service flow.
Download paper (Available to CMG members only, Click here to learn more.)
ARM-Based Performance Monitoring for the Eclipse Platform
Ashish Patel
The Eclipse Test and Performance Tools Project is releasing an open source, fully functional ARM 4.0 implementation in June 2006 (Eclipse TPTP version 4.2). This session will describe the TPTP project and explain how ARM has been implemented in TPTP. The tradeoffs involved in getting to this point will be discussed, along with the challenges in moving forward. Specific examples will show how to get started using ARM for your application with Eclipse TPTP.
Download paper (Available to CMG members only, Click here to learn more.)
A Technology Cost Model for Server Infrastructure Management
Russell A. Rogers
Today, the method used most often for budgeting server growth is linear gross estimation. The method is simple, a manager applies the percentage increase in business growth to the total number of severs they currently have in production. However this method can create significant distortions. This session reviews an approach to defining a technology cost model using principals from Activity Based Costing. The model is designed to give managers the data they need to make better decisions regarding server budgets.
Download paper (Available to CMG members only, Click here to learn more.)
Managing Financial Systems: The Peak Experience
Jon E. Schmidt
The impact of a capacity shortfall affecting financial systems can be very expensive. If a stock trading system fails to keep up with a tumultuous day, or if a retail POS system can't keep up over Christmas, the organization operating the system can lose thousands of dollars from the bottom line for each minute that the system is down or degraded. This paper discusses how some organizations track the impact of peak business demand on their servers, and how they use this information to be prepared for the next surge in demand.
Download paper (Available to CMG members only, Click here to learn more.)
Managing SOX Compliance in the Age of SOA
Hugh B. Taylor
Service-Oriented Architecture (SOA) is at the heart of many major IT initiatives and vendor offerings. However, while SOA has the potential to deliver business value through streamlined application integration, as well as integration with partners and suppliers, the open nature of SOA has the potential to cause problems for Sarbanes Oxley compliance. This session will look at compliance issues inherent in developing an SOA. Using a practical example, the paper will examine COSO Control Objectives, Risks, and their supporting IT systems from the perspective of Sarbanes Oxley compliance.
Download paper (Available to CMG members only, Click here to learn more.)
System Management by Exception, Part 6
Igor A. Trubin
Statistical Exception Detection System (SEDS) has been successfully used for more than six years to automatically produce web-based exception reports against SAS/ITRM performance data warehouse for a large, multi-platform environment. Adding some application specific metrics, including middleware traffic and response times, made SEDS an excellent tool for application performance management. This paper also describes how to create statistical control charts using a spreadsheet in order to capture a performance issue without using expensive tools such as SAS or BMC.
Download paper (Available to CMG members only, Click here to learn more.)
ARMing the Enterprise
John Yennie
Oliver E. Cole
Steve M. Sturtevant
Many large corporations and government agencies are modernizing their computer systems to use network and web-based technologies. These applications have many components and are interfaced to legacy systems as well as newer commercial, off-the-shelf packages. Response time measurements are a necessary part of the final system and ARM is a logical choice for obtaining those measurements. This session describes the ARM strategy pursued and results obtained by the e-Customs Partnership for the ACE project, a multi-billion dollar web-based customs modernization effort.
Download paper (Available to CMG members only, Click here to learn more.)
Fundamentals / Core Competency
Build a Home Computer Lab, Change Your Life and Save the Earth.
Robert D. Andresen
As IT technologies continue to change and evolve it can be difficult to keep your skills current. Training classes alone won’t help build new expertise. Unless you are able to keep working with the new technology over time, knowledge gained in a class can be quickly forgotten. Companies are selling equipment that is no longer powerful enough to run production workloads, but is quite serviceable to learn new technology. We’ll look at options for obtaining hardware and software to set up a small home network to build expertise in new technologies, to keep your current job or find a better one.
Download paper (Available to CMG members only, Click here to learn more.)
A Nifty Little Technique for Finding the Trancodes that Caused A Performance Problems
Dick Arnold
You check the weekly stats for your online systems and discover that response time, CPU usage, or whatever, suddenly increased last week. So you delve into the trancode-level stats to take a look at the big trancodes and discover that their usage went down. Now what? Well, there’s only 397 more trancodes to review, shouldn’t take too long. If you could use an easier way to isolate the cause of these sudden changes, then this easy and fairly unknown technique could be a big help.
Download paper (Available to CMG members only, Click here to learn more.)
The Bottleneck Cycle
Denise Arruda
The bottleneck cycle refers to the iterative process of finding a performance bottleneck, eliminating it, and finding the next one. The hope is that eventually the bottleneck falls outside of your application’s performance requirements. This session will present a case study to illustrate how using the principles of application performance testing enabled us to traverse our way through the bottleneck cycle. It will discuss what types of data were collected and analyzed to identify each of the bottlenecks, and discuss the approach used to eliminate each of the bottlenecks, at a high level.
Download paper (Available to CMG members only, Click here to learn more.)
Did Something Change? Using Statistical Techniques to Interpret Service and Resource Metrics.
Frank M. Bereznay
In a perfect world, one would always know the answer to this question. Unfortunately, nobody works in a perfect world. This session will explore statistical techniques used to look for deviations in metrics that are due to assignable causes, as opposed to the period to period variation that is normally present. Hypothesis Testing, Statistical Process Control, and Analysis of Variance will be explained and analyzed. SAS code to perform the analysis will be provided along with suggested reporting packages. Exploratory analysis techniques will be used to help build populations for analysis purposes.
Download paper (Available to CMG members only, Click here to learn more.)
Active Baselining in Passive Data Environments
Dr. James Bouhana
Mike Tsykin
To decide if systems are running according to their usual trend, it is necessary to compare against a performance baseline that defines an operating envelope. We describe how baselines can be derived from passive stores of performance data, which are typically flat text files or databases. Baselines can be built as needed by varying the baseline norm, granularity, update frequency, etc. Baselining outputs are sets of alert thresholds stratified by system, metric, and hour. The role and usage of baselines in automated alerting is discussed, with examples of reports that can be produced.
Download paper (Available to CMG members only, Click here to learn more.)
An Internet Business Capacity Model - More Tiers, Less Tears!
Todd R. Bourne
Mike Moroz
Theo Adis
Business is increasingly dependent upon the internet as a channel to customers. This is particularly true in finance, where the Internet can be the largest branch, turning away users for lack of resources is tantamount to closing the doors. The breadth of technologies (web, app and dbms through to traditional mainframe systems) further compounds the challenge of mapping business activity onto system utilisation. This paper reflects the real world lessons learned by CPT Global working across 16 countries building business focused capacity models to reduce costs for major financial institutions.
Download paper (Available to CMG members only, Click here to learn more.)
Traffic Capacity Testing a Web Environment With Transaction Based Tools
James F. Brady
One of the common server configurations that has emerged for supporting Web based applications is the Web Server, Application Server, and Database Server in a single frame arrangement. This paper discusses a specific traffic capacity testing experience with this server arrangement for a Windows environment using a transaction oriented load generator instead of a traditional virtual user script based tool. Both challenges encountered and insights gained are described.
Download paper (Available to CMG members only, Click here to learn more.)
New Perspectives on Benchmarking, Modeling and Monte Carlo Simulation: Operational Analysis 2.0
Dr. Jeffrey P. Buzen
By providing equations that characterize the precise trajectories of physical bodies traveling through 3 dimensional space, classical Newtonian mechanics represents a “gold standard” for performance modeling. Modern analysts study the “logical trajectories” that computer systems follow as they operate over time, traveling through multi-dimensional state spaces. With Monte Carlo simulation, random number generators are used to trace out these logical trajectories. Operational analysis provides a rationale and a set of procedures for modifying such trajectories to improve simulation accuracy.
Download paper (Available to CMG members only, Click here to learn more.)
The Straight Capacity Line
Linda J. Carroll
The trend line for a capacity forecast tends to be a boring, straight line. Most capacity reports feature this same linear regression line, year after year after year. What this author proposes is that there is a different way of developing the capacity forecast trend line to have it show the seasonal affects and/or cyclical effects while improving the accuracy of the capacity forecast.
Download paper (Available to CMG members only, Click here to learn more.)
Measuring DDF Capacity and Performance
Robert E. Chaney
The DDF (Distributed Data Facility) environment provides the capability for remote access to large DB2 databases by many remote application servers. While much of the application activity occurs outside the zseries complex, there are many measurements that can be used to understand not only the internal but also the remote activity that drives performance and capacity. This paper will explore those measurements for a large Enterprise Data Store that interacts with many various application servers.
Download paper (Available to CMG members only, Click here to learn more.)
Forecasting + Modeling: A Partnership to Predict and Prevent Capacity Bottlenecks
Margaret A. Churchill
Martha S. Hays
Capacity Management includes the monitoring of critical applications to understand when servers and service levels are close to or exceeding performance thresholds. Adding forecasting and modeling to the capacity management process allows the IT organization to anticipate when a problem will occur in the future, and prescribe the right solution, effectively preventing IT fires. This paper will compare and contrast various forecasting techniques for predicting when bottlenecks will occur. Once a prediction is made, modeling can be used to determine the impact and prescribe the right change.
Download paper (Available to CMG members only, Click here to learn more.)
CMG Italy - Best Paper: AIX Micro-Partitioning
Mark Cohen
Starting from year 2000 IBM announced the possibility to partition the pSeries family systems. Initially it was only possible to run dedicated processors in a logical partition. Now p5 series has integrated new virtual engine system technologies into their hardware and software. This introduced new features as Micro-Partitioning which provides the ability to share physical processors amoung logical partitions, Virtual Lan which provides network utilization features capabilities that permit you to prioritize traffic on shared networks and allows secure communication between logical partitions without the need of a physical network adapter, Virtual I/O which provides the ability to dedicate I/O adapters and devices to a virtual server. It allows a single physical I/O adapter to be used by multiple logical partitions on the same server. This allows consolidation of I/O servers and minimizes the amount of I/O adapters required. New concepts of shared partitioning have been implemented in the p5 series compared to the IBM mainframe. This session I will provide a detailed overview of shared partitioning and its performance metric considerations.
Download paper (Available to CMG members only, Click here to learn more.)
Citrix Benchmarks
Florin David
This session will describe the choice of workload for the benchmark, the type of tests performed, as well as a description of the tests results and their use in future capacity planning.
Download paper (Available to CMG members only, Click here to learn more.)
Creating a Software Performance Engineering Team - Lessons Learned
Gregory Dawe
''I can do that job''. Those were my closing words in the job interview with the VP of Engineering who was looking for a manager to create a software performance team. Not that I knew what I was really getting myself into. The audience for this paper is the uninitiated - the software professionals who find themselves in the arena with no prior performance experience. Lessons recounting what worked for the team’s charter, overall business approach, staffing, philosophy behind lab spending, workflow, and testing policies are discussed.
Download paper (Available to CMG members only, Click here to learn more.)
Using Native Data and Automation to Perform Rapid Triage and Reporting
Jeff Doubleday
Large amounts of system performance data can be collected using native commands. The native data can be trended to assist in capacity planning and performance management. With the automation described in this session the tasks of performance triage and capacity reporting is made less tedious. Automation allows for the creation of a repeatable process for data presentation, to enforce organizational standards, analyze multiple heterogeneous systems and save large amounts of time. Instead of managing and manipulating data in a spreadsheet time can be spent performing analysis.
Download paper (Available to CMG members only, Click here to learn more.)
Apriori Evaluation of Data and Selection of Forecasting Model
Alex Gilgur
Mike Perka
Bill Fuller
One should not have to be a statistician to evaluate whether the data are adequate and the model is good for the task at hand. The methodology used in the data evaluation tool we developed, allows us to influence the selection of the most adequate forecasting model. We shall discuss here our approach to data evaluation and how we use it for generating better models.
Download paper (Available to CMG members only, Click here to learn more.)
Quantitative Techniques to Improve Your Application Profile
Richard Gimarc
The starting point for any application performance study is to develop a profile of your application. The resulting application profile serves as the foundation for further performance tuning or capacity planning studies. This session will illustrate the use of a number of techniques from statistics and operational analysis to develop, refine, and verify the elements of your application profile.
Download paper (Available to CMG members only, Click here to learn more.)
Six Sensible Steps Towards Implementing ITIL Capacity Management
Adam Grummitt
This session describes the core approach used by a number of sites in the pragmatic implementation of Capacity Management. It describes the use of ITIL and the Capability Maturity Model to help assess current levels of implementation and strategic targets for improvement. It also discusses the use of Six Sigma techniques to improve the measures of performance. It gives guidelines for successful implementation and incorporates some of the “business as usual” practices and reports necessary to maintain the capacity management regime in terms of both performance management and capacity planning.
Download paper (Available to CMG members only, Click here to learn more.)
A Practical Approach to a Processor Migration Capacity Analysis
Robert Hamilton
After installing the latest generation mainframe there is always a need for an analysis that will determine if the latest technology is meeting capacity expectations in a regular production environment. Such an analysis could face numerous technical challenges in a rather short period of time. This paper reviews an experience with a z9 processor migration capacity analysis in a large production environment.
Download paper (Available to CMG members only, Click here to learn more.)
Bringing ITIL® to Life: Automating IT Capacity Management
Martha S. Hays
The ever growing complexity and increasing size of current IT infrastructures are the top challenges for today’s IT Capacity Managers. Facing hundreds of servers supporting vital business functions makes Capacity Management automation a must. This paper outlines how to overcome the challenges of modern large-scale environments by implementing an ITIL (IT Infrastructure Library) compliant best-practice Capacity Management methodology based on advanced analytics and business intelligence tools.
Download paper (Available to CMG members only, Click here to learn more.)
A Cohesive Framework to Quantify Computer Systems Assurance
Dr. Dominique A. Heger
Phil A. Carinhas
This study introduces a systems-engineering and evaluation methodology that focuses on the stability of an entire computing infrastructure. The conducted research elaborates on the cohesive systems assurance (CSA) methodology, which encapsulates the concepts and methods of product assurance (reliability, availability, and maintainability), performance and scalability, and dependability (security and safety). This approach substantially deviates from the pervasive systems analysis process in use today that treats the discussed dimensions individually in a vacuum.
Download paper (Available to CMG members only, Click here to learn more.)
Back of the Envelope, Rules of Thumb and Little’s Law
James Holtman
Performance analysis starts on `day 1’ of the project. There are some `rules of thumb’ that can be used to start estimating the performance and to determine a `resource budget’ for the components (how much can be spent to process a unit of work). There are some simple `back of the envelope’ calculations that can be done (and should be continuously recalculated throughout the life of the project) that will provide targets against which the software can be tested. This testing should begin as soon as the developer has a working copy for ""unit testing"".
Download paper (Available to CMG members only, Click here to learn more.)
The LOWE Down on Capacity Planning
Jim Horne
This paper will take you through the way I perform CPU capacity planning at Lowe’s Companies to forecast our continuing CPU upgrades. You will go through the SAS programs, the spreadsheet and the human interaction I use to predict what size machine Lowe’s will need and when we will need it. The SAS part will take you through collecting and summarizing the data. The Excel part will show how I use linear trending along with some human factors modification (you know, tweaking) to predict the future in six month increments.
Download paper (Available to CMG members only, Click here to learn more.)
The Myth of Memory Utilization on Midrange Systems
Brian Johnson
This session discusses the widespread use of the memory utilization metric on midrange systems and why it’s both misleading and ill-advised to report and use it.
Download paper (Available to CMG members only, Click here to learn more.)
Java Performance Analysis 301
Peter Johnson
The Java Platform provides a variety of mechanisms for monitoring the performance of Java applications. There are several tools that are freely available that can be used to monitor Java applications. This session describes some of those tools, and the Java Management Extensions (JMX) technology on which those tools are built. Additionally, the paper shows how custom tools can be built using JMX.
Download paper (Available to CMG members only, Click here to learn more.)
Getting to Know Your Production Response Time
Dr. Charles A. Letner
An important aspect of managing performance is managing response time extremes, particularly transactions with longer response times. To effectively manage the extremes requires an understanding of response time distributions. In this session response time distributions obtained from production are analyzed. The results demonstrate that response times are not normally distributed. Techniques to analyze the resulting distributions are presented. Additionally, discrete-event simulation techniques are presented that produce simulation results that approximate production distributions.
Download paper (Available to CMG members only, Click here to learn more.)
Applying Queuing Theory to Optimizing the Performance of Enterprise Software Applications
Dr. Henry H. Liu
Performance is one of the most stringent requirements for large scale enterprise software applications. It is crucial in determining the success or failure of a large project. It spans various stages of a software product life cycle from designing to developing, and to final delivering to the customer. In this paper, using the two most fundamental concepts, wait events and service demand, we demonstrate quantitatively how we can leverage well-known queuing theory to help achieve the best possible performance for large scale enterprise software applications, both efficiently and effectively.
Download paper (Available to CMG members only, Click here to learn more.)
Performance Signatures: A Qualitative Approach to Dependency Guidance
Rico Mariani
This session describes a simple qualitative approach, through approximate Performance Signatures, that allows prescriptive dependency guidance to be given in real time and facilitates improved analysis of measured results. Emphasis is placed on ease of adoption and preventing common large mistakes.
Download paper (Available to CMG members only, Click here to learn more.)
Analytic Performance Models for Single Class and Multiple Class Multithreaded Software Servers
Daniel A. Menasce
Mohamed Bennani
Modern computer systems are based on a wide variety of software servers comprised of threads that serve submitted requests. Arriving requests that find all threads busy, are placed in a queue. Threads that are busy executing requests compete for hardware resources at the machine where the software server runs. This session presents analytical models for a wide range of multithreaded software server architectures. Numerical results are presented to illustrate the use of the models.
Download paper (Available to CMG members only, Click here to learn more.)
Case Study of Modeling Performance in a Politically Charged Environment
Carol M. Petroski
Modeling performance can be extremely difficult, particularly when an outsider does the analysis. The modeling methodology and overall approach can make a significant difference to the success of such an effort. This session discusses a user experience creating a simple simulation model. This model led to the identification of a serious coding flaw in the application.
Download paper (Available to CMG members only, Click here to learn more.)
Load Testing: Points to Ponder
Alexander Podelko
Testing of multi-user applications under realistic stress loads remains the only way to ensure appropriate performance and reliability in production. The author outlines some issues to consider for performance testing of distributed business applications and presents the typical pitfalls from the practical point of view. While the original objective was to contrast load testing with functional testing, the paper touches many important points of performance testing.
Download paper (Available to CMG members only, Click here to learn more.)
Encouraging Wider Use of Performance Metrics through Web Technologies
Todd Schmitter
The IBM mainframe environment provides a wealth of performance metrics, but the exploitation of these metrics is often limited due to the expertise required to access and understand the metrics. As a result, analysis and decisions can suffer from a lack of exploiting available metrics. This session discusses the use of intranet-based technologies for greatly increasing the accessibility of performance metrics to a wider audience. It includes a consideration of the challenges of managing and presenting the data in ways that aid the analysis process.
Download paper (Available to CMG members only, Click here to learn more.)
Five Steps to Establish Software Performance Engineering
Dr. Connie U. Smith
Lloyd Williams
Most software performance problems are due to fundamental architecture or design problems. Thus they are introduced early in development, but are typically not discovered until late, when they are more difficult and costly to fix. Software Performance Engineering (SPE) is a systematic, quantitative approach for cost-effectively building performance into software systems. It consists of a process for applying the SPE methods throughout the life cycle of new systems; data required for SPE studies; software and system execution models for quantitative performance assessments...
Download paper (Available to CMG members only, Click here to learn more.)
The Roadmap for Full Lifecycle Performance Engineering
Amy C. Spellmann
Richard Gimarc
Christopher Lee
In this paper we describe and demonstrate Performance Engineering methods and techniques that apply across the software development lifecycle. This methodology, which is based on reusable tools and repeatable processes, can be applied at any stage of the lifecycle. The session includes a case study of a retail banking application where Performance Engineering was successfully applied during design, test and production.
Experiances of Using LQN and QPN tools for Performance Modelling of a J2EE Application
Nidhi Tiwari
Prabhakar Mynampati
Performance of a J2EE application is influenced by the underlying infrastructure, operating system and middleware parameters. Usually reactive approach of testing is used to configure these, which is costly and time consuming. Consequently a proactive approach of performance modeling is required. Layered Queuing Networks and Queuing Petri Nets are two such effective techniques for tuning environment. This paper articulates our experiences with these techniques for a J2EE application. The relative attributes of two techniques are listed to provide an insight on their suitability in a context.
Download paper (Available to CMG members only, Click here to learn more.)
Approach to Build Performance Model for a Web-Based System from its Application Server Logs
Suhas Sudheendra
Mitesh V. Patel
Pratik Kumar
Developers today have to ensure that their applications do not violate the Service Level Agreement at any point in time. Stringent requirements and dynamic workloads have exposed the need for predictive performance models that can be used for a ‘What-if’ analysis and Capacity Projections. However, some applications continue to be put into production without going through the required performance tests. This paper proposes an approach to build performance models of such applications without any exclusive performance testing, by extracting relevant and required data from its production logs.
Download paper (Available to CMG members only, Click here to learn more.)
Hot Topics
Grid Technology – Vision, Architecture, and Node Capacity Considerations
Phillip Carinhas
Dominique Heger
Greg Simco
The first part of this study introduces the vision of Grid networking. The study elaborates on the architecture and technologies that surround the Grid paradigm, and addresses some of the issues surrounding Grid applications. The second part elaborates on some of the challenges faced by the Grid scheduler, and introduces an analytical model that allows quantifying the capacity behavior of the Grid nodes. The study further proposes a Monte Carlo based probability estimation procedure that focuses on optimizing the communication behavior between a Grid node and the scheduler.
Download paper (Available to CMG members only, Click here to learn more.)
Security and Compliance Incident Response
Dipto Chakravarty
John Melvin Antony
This paper presents a manifesto for handling incident response related to compliance breach in corporate governance, which is an emerging software infrastructure for managing compliance and security. Public companies are failing to meet compliance because the vital cross-sections of data that is supposed to help companies meet the regulatory criteria are not readily available and what is available is insufficient to meet the regulations. This session, in a 3-part approach, introduces: a set of primitives, the concept of a “unit of work”, and characterizes its incident response workload.
Download paper (Available to CMG members only, Click here to learn more.)
Utilization is Virtually Useless as a Metric!
Adrian Cockcroft
We have all been conditioned over the years to use utilization or percent busy as the primary metric for capacity planning. Unfortunately, with increasing use of CPU virtualization and sophisticated CPU optimization techniques such as hyperthreading and power management the measurements we get from the systems are virtually useless. This paper will explain many of the ways in which the data we depend upon is distorted, and proposes that we turn to direct measurement of the fundamental alternatives and express capacity in terms of headroom, in units of throughput within a response time limit.
Download paper (Available to CMG members only, Click here to learn more.)
On the Number of Partitions
Yiping Ding
Virtualization is still today the hot topic it's been for several years. It has entered production environment both for server consolidation and for the rollout of new applications. Several implementations are available to choose from. In this session we address one of the first questions you must ask: how many virtual systems can comfortably run on a particular physical system, with known processor, memory and disk resources? We introduce a statistical model that helps you answer this important question.
Download paper (Available to CMG members only, Click here to learn more.)
Ever Feel As If The World Is Passing You By? Wanna Catch Up Fast?
Dr. Bernard Domanski
Rob Domanski
PDAs, Webcams, RSS feeds, podCasting, Windows Mobile, iPods, Torrents, RAR Compression, digital movies, mp3s, mp4s ... not exactly the buzzwords we were using 5 years ago. This entertaining session will focus on these and other 'hot’ technologies. Hopefully, you’ll walk away with knowledge to apply back at the office ... or at the very least, some cool toys to add to your Christmas gift list!
Download paper (Available to CMG members only, Click here to learn more.)
The Virtualization Spectrum from Hyperthreads to GRIDs
Dr. Neil J. Gunther
Virtualized systems remain opaque because they have surpassed the measurement paradigms of most performance management tools. We ameliorate that problem by (1) recognizing that the disparate forms of virtualization lie on a discrete spectrum bounded by hyperthreading at one extreme and GRID-like services at the other, (2) observing that the YGJPM TY frequency determines the relative position and performance of each system on this VM-spectrum. Several examples illustrate how this new insight can make each VM-spectral region more visible for performance management.
Download paper (Available to CMG members only, Click here to learn more.)
ITIL vs. Agile Programming: Is The Agile Programming Discipline Compatible With The ITIL Framework?
Charles Hoover
Information Technology Infrastructure Library (ITIL) is an approach to IT Service Management that is being widely accepted across the industry. Agile programming methodologies like Extreme Programming (XP) are new ways of developing software. The question is: Can these two processes work together in the same organization successfully?
Download paper (Available to CMG members only, Click here to learn more.)
Measuring and Modeling the Performance of the Xen VMM
Jie Lu
Lev Makhlis
Jianjiun R. Chen
Server virtualization technology provides an alternative for server consolidation by creating a set of logical resources that share underlying physical resources. Xen virtual machine monitor, a popular virtualization solution in Linux world, supports execution of multiple guest operating systems with unprecedented levels of performance and resource isolation. Performance modeling of virtual servers faces challenges of obtaining meaningful measures as the operating system deals with virtual resources. This paper presents a practical approach for measuring and modeling the performance of Xen.
Download paper (Available to CMG members only, Click here to learn more.)
Targeted Capacity Planning: Delivering a Business Focussed, Cost Effective Service.
Rod Parsons
In today's ITIL based world the focus of IT Service Delivery is on the management and support of services to its customers: The business. These services, primarily based around applications designed to deliver products to external customers, have widely varying IT infrastructure requirements and business value. Conversely, the Capacity Planning group was predominately organised along architectural platform lines, focussing separately on mainframe, midrange Unix and Linux, Windows, network and storage architectures. While creating a business-focused capacity plan for an application using a single architecture was, in theory, simple under this organisational structure, complex, multi-platform applications, however, were not readily supported. To fully understand the capacity impact of changes in business demand on these applications it was necessary to build an end-to-end view of the application and to report capacity utilisation across the relevant platforms. Support for both platform based and end-to-end capacity planning clearly required a different organisational framework. This presentation covers the development of a framework with a multi-tier approach to delivering its capacity monitoring, reporting, analysis and forecasting service. It also describes our journey towards implementing end-to-end capacity planning under this framework.
Download paper (Available to CMG members only, Click here to learn more.)
Optimization with Service Level Objectives in Virtual Environment
Dr. Anatoliy Rikun
Yiping Ding
In this paper we analyze important problem of distributing a set of applications among physical servers. Taking into account the factor that different applications may have different peak times, different resource consumption patterns, and different importance levels may significantly improve overall system performance and QoS. This paper presents a mathematical model and some case-study examples for this problem. We also discuss some optimization aspects of tuning the applications parameters after their deployments.
Download paper (Available to CMG members only, Click here to learn more.)
Performance Tuning and Resource Management in Java Applications
Robert E. Ritchie
More and more mission critical and large scale applications are now running on Java. With the rapid increase in market share of application servers like Websphere and Weblogic, performance management of Java applications continues to be a challenge. As Java applications continue to evolve, tuning performance and managing server and network resources becomes even more overwhelming. This session explores high level philosophies for performing these tasks throughout the development life cycle and for pro-actively monitoring Java applications in a production environment.
Download paper (Available to CMG members only, Click here to learn more.)
It May Be Virtual, ... But the Overhead Isn't
Dr. Michael A. Salsburg
Peter Karnazes
William Maimone
This paper presents the results of benchmarks that were designed to measure and observe the performance behavior of server virtualization. Results are published for both Xen and VMWare virtualization technologies. It includes a study of the effects of various I/O and CPU intensities. The benchmarks were designed to observe the overhead of virtualization and the affects of technology and workload mixes on that overhead.
Download paper (Available to CMG members only, Click here to learn more.)
The Well-Managed Web Service
Herb Van Hook
This paper discusses successful operational management of the emergent technologies represented by adopting Web Services for application integration. It is important to recognize that the next few years will be marked by significant trends in the packaged and custom application market. These include: - Alignment, convergence and consolidation among application logic - Application integration and application servers - Adoption of service-oriented architectures - Adoption of Web Services - Increased use of programmable business process languages - Integrated and complete management of this environment
Download paper (Available to CMG members only, Click here to learn more.)
Real World Adventures in Server Virtualization
Peter J. Weilnau
Stephen Marksamer
Server Virtualization continues to be one of the hottest topics of 2006. Join our quest for enlightenment as we explore the performance data of a large enterprise environment. We’ll look closely at data captured from VMware ESX, its virtual machines and Windows servers seeking correlations between activity levels, response times and other key measurements both before and after virtualization. When our quest for enlightenment is complete, we’ll provide recommendations to aid in your virtualization endeavors.
Download paper (Available to CMG members only, Click here to learn more.)
Network / Internet
Achieving Practical Network Application Impact and Response Time Projections
James H. Baxter
Preparing for the successful deployment or expansion of a networked application that offers acceptable performance at remote locations often involves more than just ensuring adequate network bandwidth. This paper outlines the concepts, tools, and processes utilized to accomplish a practical network impact assessment that includes end-user response time projections. Elements of these techniques can also be employed to reveal the real source of poor application performance and eliminate finger-pointing between network and application support personnel.
Download paper (Available to CMG members only, Click here to learn more.)
Measurement of Transaction-Based End-To-End Response Time in Un-Armed Environments
James Bouhana
Mike Tsykin
Christofer D. Langshaw
Transaction-based response time is the key metric for assessment of SLA compliance, installation accounting and performance trouble-shooting. Both end-to-end measurement and breakdown by components of transaction path are required. This session reviews the available methods and focuses on environments that are not instrumented with ARM - the vast majority of installations. Theory and available tools are reviewed and a successful implementation is discussed in detail.
Download paper (Available to CMG members only, Click here to learn more.)
Identifying Network Failures and Evaluating Link MTBF from Utilization Logs
Paolo Cremonesi
Dr. Guiliano Casale
Stefano Visconti
Network failure detection techniques and link Mean Time Between Failure (MTBF) estimates are required to assess the reliability of large communication networks. We present an experience based on utilization logs of a large ISP network comprising hundreds of links, and spreading over a geographic area. The complexity of the network requires to account also for the mutual dependencies between events on different links. Nevertheless, we show that robust non-parametric data mining methods offer a simple and effective way to accomplish the task.
Download paper (Available to CMG members only, Click here to learn more.)
Evaluation and Comparison of Search Engines Using the LSP Method
Dr. Jozo J. Dujmovic
Haishi Bai
We present a comprehensive model for quantitative evaluation and comparison of search engines. The model is based on the LSP method for system evaluation. The basic contribution of our approach is the aggregation of all relevant attributes that reflect functionality, usability, and performance of search engines. In this respect our model is fully consistent with the ISO 9126 standard for software product evaluation. Performance analysis of search engines is based on our search engines benchmarking tool that is also described in the paper.
Download paper (Available to CMG members only, Click here to learn more.)
Ten Commandments of TCP/IP Performance
Nalini J. Elkins
The ten commandments of TCP/IP performance are a distillation of hard-won experience. Monitoring and tuning TCP networks on the mainframe is complex for the basic reason that each network is a mixture of many applications and pieces of hardware. Each connection contains layers of protocols and subprotocols which must be decoded to make sense of the traffic patterns. Making sense of it all is the first step to tuning and improving performance.
Download paper (Available to CMG members only, Click here to learn more.)
A Tutorial on SIP Application Server Performance and Benchmarking
Dr. Curtis Hrischuk
The Session Initiation Protocol (SIP) is an internet protocol for establishing sessions between two or more parties. It is becoming ubiquitous in uses ranging from Voice over IP, instant messaging, Internet TV, and others. The Java community has even provided a standardized API so that SIP applications can now be built within J2EE application servers. These new capabilities also bring with them new performance engineering methods, tools, and benchmarking needs. This paper describes the experiences and processes for the performance engineering of SIP applications in a J2EE environment
Download paper (Available to CMG members only, Click here to learn more.)
Instrumentation and Analysis of Web Transactions in a Large Multi-Tier Banking Services Application
Mark W. Johnson
Bret Patterson
Transactions in mission-critical applications are monitored and analyzed to avoid performance and availability problems that impact end-users, and to tune application and resource performance. This session explores a large European banking service’s requirements for managing the transactions of a new multi-tier web application, the approach taken to address the requirements, and the technical solution employed for instrumentation, monitoring, and reporting.
Download paper (Available to CMG members only, Click here to learn more.)
10 Steps to Securing Your Web Applications
Peter Johnson
You set up a firewall, run anti-virus and anti-spyware software, and regularly patch security holes in your operating system. Therefore your data is secure, right? Wrong! Sloppy programming practices can leave your data open to prying eyes. This paper describes the top 10 security vulnerabilities that can be found in today's web-based applications, and how those applications can be changed to avoid these security issues. The paper also provides a brief survey of commercial products that claim to catch several of these vulnerabilities without having to change the applications.
Download paper (Available to CMG members only, Click here to learn more.)
Softswitch Testing
Garland Kan
Without testing the true real world capacity, performance, and voice quality of a Softswitch, these metrics cannot be quantified. Testing Softswitches is new to the industry. Softswitches has only been around around 5 years and new techniques has to be employed find the capacity, performance, and voice quality it will provide. Using various tests, all of these metrics can be found and a one to one comparison between Softswitches can be made using the results. This session will go into details on what types of tests to perform and what results will come out of those test and what they mean.
Download paper (Available to CMG members only, Click here to learn more.)
Storage
Workload Characterization Algorithms for Remote Copy
Dr. H. Pat Artis
The planning and implementation of synchronous and asynchronous z/OS remote copy solutions present a myriad of workload characterization problems as well as risks to the enterprise. This presentation will discuss algorithms for calculating write data and I/O rates, identifying problem volumes, and present a methodology for testing and certifying a remote copy environment before committing your production workload.
Download paper (Available to CMG members only, Click here to learn more.)
Understanding the Performance Implications of MIDAWs
Dr. H. Pat Artis
Modified Indirect Data Address Words (MIDAWs) are a new feature of z9 processors, enabled by z/OS 1.7, that can substantially improve the performance of FICON connected storage subsystems. This session will provide an overview of MIDAWs and discuss how they can be employed to restructure complex channel programs. After this introduction, the session will focus on the analysis of measurement results for a number of structured experiments designed to demonstrate the performance benefits of MIDAWs.
Download paper (Available to CMG members only, Click here to learn more.)
Database Backups Using Virtual Tape Volumes
Kathleen N. Hodge
The time it takes for the database administrator to conduct the database recovery process directly affects the company’s bottom line. Database downtime impact can be measured in billions of dollars resulting in lost opportunity and disgruntled customers. This paper describes virtual tape technology and advantages which can be realized by the database administrator when database recovery efficiency is essential. Virtual tape offers an alternative to expensive disk storage and physical tape cartridge dependencies.
Download paper (Available to CMG members only, Click here to learn more.)
Cache Management of Competing I/O Workloads
Bruce McNutt
Recent developments in the SCSI-3 standard make it possible for a storage system to develop performance management strategies, based in part, upon the relative importance of competing applications. The compelling benefits of such management, applied to physical disks, were the subject of a companion paper, presented one year ago. This paper extends this discussion to another key storage system resource: the cache memory. We present a strategy for such management, and illustrate its potential advantages.
Download paper (Available to CMG members only, Click here to learn more.)
Remote Copy 100 km testing
Bruce McNutt
Measurements were done to look at the performance impact of synchronous storage remote copy as distances increase from 0 to 100 km. Three environments were looked at: Single site only workload, with DB2 GBP CF structure duplexing, and with dual site workload. This summarizes the results seen. It is also published on the IBM web site.
Download paper (Available to CMG members only, Click here to learn more.)
Bertha: A Benchmark Tool for High-Performance Storage Subsystems
Lewis Myers
Dave Wagoner
John Lenehan
The popular I/O benchmark tool “Bonnie” has been modified to form a distinct benchmark tool called “Bertha”. The features of Bertha that distinguish it from Bonnie, as well as most other benchmark tools, are 1) it generates substantially more I/O to better measure capabilities of contemporary storage subsystems, 2) it offers a unique facility to “replay” I/O, “transactions” from actual production applications or simulations, to more accurately suggest performance that will be experienced when using particular applications, and 3) it providesextensive metric reporting that provides considerable insight.
Download paper (Available to CMG members only, Click here to learn more.)
Forecasting Database Disk Space Requirements: A Poor Man's Approach
Edward L. Tretel
In the absence of special purpose monitoring and/or modeling software designed specifically for forecasting database disk space requirements, a solution was developed using general purpose database facilities and office suite products. The results achieved were (1) an understanding of the heretofore unknown trends and patterns in the use of disk space by individual databases (2) the ability to accurately and proactively forecast the additional disk space needed for individual databases, and (3) the ability to reclaim the forecast unused disk space, all based upon linear regression analyses.
Download paper (Available to CMG members only, Click here to learn more.)
Can You Afford Low Cost Storage?
James A. Yaple
Greg Lee
James Barton
With discount computer stores selling disk drives for a dollar per GB, how can enterprise storage systems be more than 100 times that? What are the advantages of enterprise storage arrays? What role can low cost storage play in the information life cycle? Initiatives from companies such as Oracle are promoting lower cost storage alternatives in departmental and enterprise environments. This session examines these questions and present performance results of various arrays using the ORION benchmarking tool.
Download paper (Available to CMG members only, Click here to learn more.)
Unix / Linux
Capacity Planning by Simulating UNIX Servers
Uriel J. Carrasquilla
This session presents the development and testing of simulation software for the measurement, and modeling of UNIX performance data for data centers with mission critical UNIX servers.
Download paper (Available to CMG members only, Click here to learn more.)
Out-of-the-box Performance of OLTP on High-end Servers - A Comparison of File Systems & Configuration
Timothy P. Cook
Discussion of results of a study into a high-end OLTP workload running on Oracle 10g on top of today's popular filesystems - UFS, VxFS, QFS and Oracle's ASM. What was seen from an out-of-the-box configuration, and what tunings were seen as beneficial and why.
Download paper (Available to CMG members only, Click here to learn more.)
AIX System Performance Experiences and Basic Tuning
Irvin G. Eiceman
Recent conversion from a Mainframe environment to AIX/UNIX presented many interesting challenges. Fortunately, many of the fundamental performance tuning principles learned in the mainframe environment applied well in the sense that you are still balancing CPU memory and I/O. As such, much of this paper is derived from personal experiences in extensive tuning of AIX systems. The intent of this paper is to help make the transition from MAINFRAME to AIX/UNIX a bit friendlier for the Performance Specialist.
Download paper (Available to CMG members only, Click here to learn more.)
The Need for Speed: Simple Tested Techniques to Beef Up Performance of Your Solaris/Oracle Database
Peg McMahon
Bob Sneed
When their C-Level execs are calling your C-Level execs about performance problems with your application, it’s time to act. The app needs an extreme makeover - but there’s no time or money for a re-write. Instead, we took it to the lab. In about two weeks of testing, we reduced CPU utilization on the database more than 90 percent and cut transaction time in half. A lot of the problem turned out to be the file system. It was not write concurrent. This is not rocket science. If your Solaris/Oracle database needs fast response under high transactional loads, here is some news you can use.
Download paper (Available to CMG members only, Click here to learn more.)
Windows
The Reality of Virtualization for Windows Servers
Mark B. Friedman
This paper discusses the performance and capacity concerns that arise when Windows servers are run as virtual machine guests on current virtualization solutions. It reviews the advantages and disadvantages of virtualization as a server consolidation strategy. It describes the major sources of performance degradation that applications running on guest machines face today and discusses the prospects to resolve these problems as new hardware emerges in the near future.
Download paper (Available to CMG members only, Click here to learn more.)
Core System Event Analysis on Windows Vista
Dr. Insung Park
Event Tracing for Windows (ETW) has been the key instrumentation technology on Windows platforms for years. Many core operating system components have been instrumented with ETW, providing a basis for system activity analysis and problem diagnosis for developers and tools. The upcoming Windows Vista® operating system contains many new events, in response to the growing need to diagnose and tune various system and application activities. We describe the system events that are available on Vista and provide a few correlation techniques that can be used to analyze them.
Download paper (Available to CMG members only, Click here to learn more.)
Utilizing Performance Monitor Counters to Effectively Guide Windows and SQL Server Tuning Efforts
Jeffry A. Schwartz
Windows and the SQL Server database management system provide many performance counters. However, the vast number of counters, and the fact that numerous explanations provided via the Windows Performance Monitor simply restate the name of the counter, make resolving performance problems in this environment a daunting task. This paper discusses methodologies, metrics, and techniques that can be utilized during a performance crisis to ascertain the cause(s) of the problem and determine an appropriate solution.
Download paper (Available to CMG members only, Click here to learn more.)
zSeries
Measurement and Modeling of DB2 zIIP Workloads
Ned A. Diehl
IBM System z9 Integrated Information Processors (IBM zIIP) provide very attractive price / performance options for appropriate DB2 environments. They also add complexity to performance management and capacity planning. This paper will discuss data sources that can be used to estimate the amount of current DB2 workload eligible for zIIP processors. Modeling techniques using that data will show approaches to evaluating the benefits of various zIIP processor configurations. Also discussed will be some of the data changes that will result from implementation. DDF will be the primary focus.
Download paper (Available to CMG members only, Click here to learn more.)
Designing and Managing FICON Inter-Switch Link Infrastructures
Stephen R. Guendert
H. Pat Artis
FICON Inter-Switch Links (ISLs) provide an elegant solution to the bandwidth requirements of complex parallel sysplex environments. This session will review basic ISL concepts, how they are defined in HCD, focus on design principals for fault tolerance, and examine the measurement data available for performance management and capacity planning.
Download paper (Available to CMG members only, Click here to learn more.)
CICS Open Transaction Environment And Other TCB Performance Considerations
Steven R. Hackenberg
With the introduction of the CICS Open Transaction Environment (OTE) architecture introduced with CICS Transaction Server 1.3, we now have much more flexibility for CICS to exploit more CPU engines without the need to split a single region into multiple regions, with the added benefit of reducing CPU consumed per transaction. The implementation of this architecture is fairly simple, but improper specifications can have ramifications which diminish or even hamper the advantages of the new approach.
Download paper (Available to CMG members only, Click here to learn more.)
The Effect of Distribution and Correlation Statistics on the DB2 Optimizer
Tom Moulder
Ever wondered about DB2 performance? Why some statements work well and not others? This presentation will explain the difference between DB2 without help and DB2 with help. The particular help that will be discussed is the ability of DB2 to collect statistics concerning data content that can be used to improve performance. Don't let the title scare you. Nothing in the presentation will ever use anything more that addition, subtraction, multiplication and division. The longest calculation is of four values. Simple Math can be used to unlock the secrets of DB2.
Download paper (Available to CMG members only, Click here to learn more.)
Dials for an PM Dashboard: Velocity’s Missing Twin, and Quantifying Surprise
Rich Olcott
This session will discuss what should and should not be in a PM dashboard, what is missing from the z/OS displays we have now, and an efficient, platform-agnostic way to track and highlight what is happening in our systems.
Download paper (Available to CMG members only, Click here to learn more.)
Use Trending to Manage Application and System Performance
Bruce Perkinson
In the IT world we are all feeling the effects of some form of cost reduction. While companies may be successful in reducing staff costs they are also sacrificing legacy knowledge and technical expertise. This makes simplifying Performance Management more critical than ever. By collecting and trending a few key statistics, we can monitor the performance and identify important trends for everything from batch cycle elapsed times to Netview CPU consumption. We will review how trending helped identify opportunities to reduce CPU consumption, shorten batch run times, and reduce cost of ownership.
Download paper (Available to CMG members only, Click here to learn more.)
A Performance Analyst’s Guide to the RMF Type 70 Record
William L. Shelden, Jr., Ph.D.
The evolution of the performance data in the RMF type 70 record is described from the earliest MVS systems running on uniprocessors to current z/OS systems running in logically partitioned systems, on CECs with various types of CPs, in LPARs with combinations of shared and dedicated logical processors under the control of Workload Manager using the functionality of the Intelligent Resource Director. Techniques for analyzing and reporting on the metrics in the RMF type 70 record will be discussed along with techniques for using the data to address issues of CPU contention and latent demand.
Download paper (Available to CMG members only, Click here to learn more.)
Effect of Parallel Access Volumes (PAV) Technology on z/VM Guest Disk I/O Performance
Dr. Brian K. Wade
In May 2006 IBM equipped its z/VM operating system to exploit IBM storage servers' Parallel Access Volumes (PAV) technology, so as to expedite virtual machine disk I/O. This session briefly describes how z/VM uses PAV. It also gives the results of measurements illustrating the effect of the PAV exploitation on virtual machine disk I/O performance. The measurements show that z/VM's use of PAV can provide performance improvements for virtual machine disk I/O.
Download paper (Available to CMG members only, Click here to learn more.)
Business Performance Management |
Fundamentals / Core Competency
Hot Topics |
Network / Internet |
Storage |
Unix / Linux |
Windows |
zSeries
Home |
Conference |
Groups |
National |
Members |
Links |
Site Map
|
