| home | about | membership | TEC 2009 Event | worldwide CMG | service directory | | |
|
|
|
Official Media Partners 
|
AbstractsMONDAY 26th June 2006Session 1A1 - Track A - MAINFRAMEMigrating to z-990 - A user Experience
Awaiting Abstract Session 1A2Much Ado About CPU - Now with z9 IBM mainframe processors have in recent years introduced a number of capabilities of real value to customers. These capabilities have, however, required changes in the way we think about CPU management. This presentation describes these capabilities and how to evolve your CPU management to take them into account. It is based on the author's experience of evolving his reporting to support these changes, including the new z9-109 processor. Session 1A3Modelling zSeries processors How fast is your z890, z990, or z9 processor? How much faster (or slower) will your next planned processor be? zSeries processors run at different speeds for different workloads. zSeries processors deliver different MIPS values for different configurations. Is fewer faster engines always good? The presenter uses the IBM zPCR tool together with an analytical modelling tool, to walk through a case study that addresses the variability in performance and capacity, of the zSeries processors. We look at issues including PR/SM overhead, effective dispatch time, logical CP versus physical CP, increased TCB time, and LSPR ITR ratios. As well as examining the effects of these issues, the presenter uses an analytical model to predict workload response time when workloads are migrated to a new zSeries box. Session 1A4MIDAWs - Go Faster Stripes for DB2? Martin Packer will present a simplified form of Jeff Berger's paper on the new z9-109 MIDAW facility and how it can speed up DB2 access to data. Jeff is a key performance designer in the DB2 for z/OS Development Laboratory. Included is some recent information on FICON and DS8000, together with how DB2 takes advantage of the z9-109 MIDAW facility. Session 1A5
Swimming the DB2 Pools : Understanding and Tuning DB2 Storage Awaiting abstract Session 1A6
Top Ten Best Practices for Improved z/OS Performance and Lower TCO Yes, it is possible to improve performance of a z/OS environment while also lowering the TCO (Total Cost of Ownership). The areas covered in this session include: CICS, DB2, IRD, PR/SM, VWLC, WLM, zAAP, just to name a few. This Top-10 collection is based on recent experience from over 50 major installations. Attendees of this session will learn proven best practices on how to set up, customize, report, and analyze the performance and capacity of z/OS and its major subsystems while never loosing sight of the effects on the TCO. TRACK B - DISTRIBUTEDSession 1B1Monitoring performance in a virtual infrastructure
This presentation discusses the capture and analysis of performance data in
a virtual infrastructure. It focuses on two widely-used architectures: VMware
ESX Server/Virtual Center and Microsoft Virtual Server. Reference will also
be made to Solaris Containers and Linux Xen. Topics covered will include:
Session 1B2Capacity planning considerations for Windows virtual machine technology
Awaiting abstract Session 1B3Performance Benchmarking; Updating the Toolset Filesystem We draw on experience gained in competitive benchmarking of Solaris ZFS against competitor products from Veritas, Microsoft and the Linux community. Rather than dealing with the specific results of the testing, we survey the tools that have been used "traditionally" through habit or for reasons of availability and examine their strengths and weaknesses in the face of current trends in data storage and access. We present FileBench - an extensible, open source filesytem benchmarking framework which provides more realistic load generation. This allows for emulating existing load-generation tools or for the simulation of the data access patterns of current applications. Session 1B4Pre-Deployment Sizing Made Easy: Exchange 2003 Case Study Estimating the necessary hardware resources when deploying Exchange is more art than science. Typically, field experts rely on past experience and conjectures to recommend a deployment architecture that meets the requirements. Microsoft System Center Capacity Planner (SCCP) eliminates this guesswork by guiding users through the process of sizing and architecting future deployment of highly geo-scaled distributed applications such as Exchange 2003. Users can fine tune recommendation by changing the workload characteristics or hardware configuration with the use of a graphical topology editor. A sophisticated simulation predicts the hardware utilizations and transaction latencies which allows the user to troubleshoot potential performance bottlenecks. This paper will give an overview of the technology and focus on the process used to validate SCCP's topology recommendations and how the accuracy of the simulation predictions was verified for such a complex application. Session 1B5Challenges of Workload Characterization in Parallel Processing
Environments SQL response time in parallel processing environments is sensitive to the degree of parallelism within the processing of individual SQL queries and the manner in which SQL access patterns are distributed across data base tables. These two considerations, which do not arise in conventional systems, create special challenges for workload characterization. The situation is further complicated by the fact the mechanisms for regulating intra-SQL parallelism differ materially among Oracle 10g, DB2 UDB ESE and Teradata. This paper examines the workload characterization issues that arise in each of these architectures, and includes a discussion of data sources, analysis procedures, and applications. Session 1B6 Distributed TrackExperiences in Capacity Management of Shared UNIX Infrastructure This paper looks at some of the experiences at the National Australia Group (Europe) with the Capacity Management of a large distributed shared infrastructure running on Sun Solaris. In particular, the paper focuses on our Oracle and Websphere infrastructures, both of which are supporting multiple applications, services and business units on the same sets of hardware. The challenges faced include setting up effective monitoring, data collection, workload classification and the determination of some new methodologies for the Capacity Management of multiple applications running on horizontally scaled UNIX hosts. TRACK C - WEBSession 1C1An Internet Business Capacity model - more tiers, less tears!
Business is increasingly dependent upon the internet as a delivery channel to customers. This is particularly true in a banking environment, where the internet channel is effectively the largest branch, and turning away users because of insufficient capacity is tantamount to closing the doors. From a capacity modelling perspective, the breadth of technologies involved - web, app servers, db servers through to traditional mainframe systems - further compounds the challenge of mapping business activity onto system utilisation. This paper reflects the real world lessons learnt by CPT Global consultants working across 16 countries building business focused capacity models and reducing costs for major financial institutions. Session 1C2Application Behaviour in a Multi-platform world The days of having a single platform for each critical business applications are long gone. Today's business climate demands a speed of development and deployment that often means having application components built quickly with this likes of Java and then spread across different platforms. This is great - until something goes wrong. This session will look at a new generation of tools specifically designed to help organisations investigate the causes and challenges of multi-platform Java application problems - while at the same time integrating with existing applications and tools that the business already relies completely upon. Session 1C3More Performance, More Scalability for your Java Apps with Less Hardware
The presentation would cover the following: Session 1C4Through the Prism of Fractals: Why SOA Should Reflect the Natural
Order Service Orientation is emerging as the fourth wave of the computing paradigm shift because it promises to enable broad-scale interoperability and unprecedented business agility in a service value-net (ecosystem). Containing IT complexity and aligning IT with business through a set of sound and robust design principles are pivotal to the transformational power of Service Oriented Architecture (SOA). This paper looks for insights into containing IT complexity by studying the time-tested tenets and dynamics of complex fractal-like forms that abound in Nature. Session 1C5 Management TrackFight the Symptom or Fix the Cause - Evolving from Resource to
Service Oriented Capacity Management Traditional capacity techniques have often focused on symptom alleviation - identifying trouble spots and specifying which resources need to be augmented or replaced to meet demand. Such techniques can lead to a detailed understanding of the usage of individual resources, but this is precisely - and perhaps counter-intuitively - the problem. Organizations may continually add resources with little or no improvement to the overall end user experience. Modern distributed systems require a different approach. Service-Oriented Capacity Management is a logical evolution from the trend of managing IT as a service. This practice focuses on fixing the causes of poor performance and the impact on the end-user experience by encouraging management of the infrastructure focused on business needs. This session will discuss the nature of SOCM in more detail. Session 1C6Capacity Managing SOA. A whole new ball game? Friends Provident are in the process of implementing SOA. I have been investigating the potential impact of SOA on the roles of Capacity Management and Monitoring. This presentation will go through my findings and highlight the challenges that we will face through the introduction of SOA. Is it a whole new ball game for us? I'll let you be the judge of that. TRACK D - GENERAL (Beginners)Session 1D1 & 1D2System architecture and performance Part 1 and Part 2
A good understanding of the way a whole system works is essential to the design
and implementation of reliable, efficient and scalable application software.
Overall system performance is determined by the underlying hardware configuration
and by how the operating system manages the hardware resources and makes them
available to the application software. This seminar will provide a technical
overview of the way that the hardware and operating system elements within
a computer system function and interact with each other. Session 1D3 General TrackWhy We Created UKCMG A light hearted look at the early years but outlining the serious reasons for setting up the group and defining its objectives (which are still the same). Session 1D4 General TrackAny Why Nobody Seems to Have Noticed The second session will try to explain why the same old problems are still there despite all the technological developments that have occurred and despite the number of papers that have been written and presented on the topics of Performance Management, Capacity Management, Service Management and all the other UKCMG featured topics. Session 1D5A Ten-Step Plan to Effective Capacity Management Designing and building an effective capacity management function has never been harder: the competition for talented staff is high, as is the current workload of projects and 'business-as-usual'. At the same time salary and training budgets are often constrained. How does the contemporary capacity manager cope with these challenges? Follow our ten-step plan to designing and building an effective capacity management function; wherever your starting point and whatever your challenges you will find useful techniques here! Session 1D6Is there anybody out there? Capacity Planning in the real world
UKCMG is all about Performance and Capacity Management, but does anyone actually do it in the real world? This paper proves with actual examples that there are companies that do regular Performance and Capacity Planning and that they gain real, but not always obvious benefits from doing so. TUESDAY 27th JUNE 2006Session 2A1 MainframeAre your systems SOX and COBIT compliant?
The presentation highlights the weaknesses in the z/OS (MVS and USS) systems
areas that are addressed by the Sarbanes Oxley Act and COBIT requirements.
Further we aim to show how the products eventACTION and ussACTION help enforce
these requirements and thereby reduce the risk of auditing exposures and possible
penalties due under this act. Session 2A2 Mainframe
Performance Architecture and Performance Management within BT
Whilst numerous presentations over the years have dwelled on a whole range
of Performance Management and Capacity Planning related topics , very few
have focused on the mission critical area of 'Performance Architecture' -
a discipline of considerable breadth which I loosely define as :- Session 2A3 Mainframez/OS WLM - Sysplex Management When it comes to z/OS WLM management within a Sysplex, I've always likened a 4 system Sysplex to a quartet. Each musician is playing their own instrument, but listening to the other musicians to ensure they are always in harmony. For what ever reason, more and more installations are consolidating systems into fewer numbers of Sysplexes. The result is more and more Sysplexes containing unlike systems with unlike or mutually exclusive workloads that need to be managed towards different goals and importance levels. Yet each Sysplex as a single WLM service definition to which all workloads must be managed. So think of this as two quartets playing side by side to different tunes. Come to this presentation if you want to hear Peter Enrico discuss WLM Sysplex management. Session 2A4 Mainframe - Application Performance TrackQuantifying benefits of zOS Performance Tuning Over the last three years, John Lewis have implemented many zOS performance and capacity management innovations. As a performance analyst I know these have been worthwhile; but given an archaic and misleading chargeback system, and continuing mainframe upgrades, how do I convince business management that things have improved? This paper explains why it can be difficult to justify performance work. It then describes an approach I've taken to quantifying the financial benefits of tuning changes sufficient to convince the most cynical manager. In short, this is how I justify my existence! Session 2A5 Mainframez/OS Performance on the New Generation of Itanium 2 Mainframes
Platform Solutions, Inc. (PSI), is the first developer of a new generation of mainframe computers compatible with the broadest set of datacenter environments and operating systems, including IBM(r) z/OS(r). The PSI offering is based on Intel Itanium 2 technology. This presentation describes the architectural and performance characteristics of the PSI mainframe. Instruction set architecture realization, that is, how z/Architecture(tm) is achieved on Itanium, is discussed. Architectural characteristics of Itanium that make it ideal for this application are described. Implementation of real and virtual I/O subsystems is also covered. Session 2A6 MainframeMXG Update Awaiting abstract Session 2B1 Distributed TrackITIL Capacity Management -- More than Charts over Coffee Many organizations are embracing ITIL as a model for best practices. While managers tend to give most of the initial attention to the Service Desk and to Service Support processes, the same managers are struggling with giving the proper focus to Service Delivery processes, Capacity Management in particular. This paper provides a view of a few organizations in varied lines of business and discusses capacity management as practiced in those environments. Those practices are then compared and contrasted with Capacity Management as covered in the ITIL Service Delivery book along with some recommendations for incorporating best practices in any environment. Session 2B2 Distributed TrackUSCMG Best paper - ITIL Capacity Management Deep Dive
ITIL is continuing to grow in acceptance in IT environments as a model for
best practices. This paper provides a low level analysis of the ITIL capacity
management discipline. The paper describes the differences between business,
service, and resource capacity management, and the need for each to have a
proactive capacity management process. The paper will describe the elements
needed for an ITIL-based capacity plan, a discussion on what several companies
have done with ITIL capacity management, and lessons learned from implementing
ITIL capacity management discipline in their environment. Session 2B3 Distributed TrackBridging the Gap Between Business and IT: Decisions IT People
should not make
This is an interactive workshop as well as presentation. It has the following
sessions: Session 2B4 Distributed TrackThe IT Bermuda Triangle?
The three IT disciplines SLM, Financial Management and Capacity Management
form the core Service Delivery within ITIL Service Management. Collectively
and when implemented correctly, these IT areas can contribute to an effective
and thus successful IT Organization. Paradoxically these areas also form some
of the biggest mysteries of modern IT times. Session 2B5 Distributed Track'Something Wicked This Way Comes': Unlocking Office Politics With
Shakespeare
Jean-Paul Sartre once observed that 'Hell is other people'. Unfortunately
we all work with 'other people'. The BBC series 'The Office' captured the
diverse personalities of people who inhabit work places in excruciatingly
accurate detail, but most people don't realise that Shakespeare was creating
his own version of 'The Office' 400 years before David Brent hit our screens.
Managers, advisors, consultants, communicators, political operators, employees
and employers, tradesmen and professionals, each with their own foibles and
agendas, all stroll across Shakespeare's stage. Each of them has things to
say about what to do, and more importantly what not to do, in the business
of dealing with other people. Session 2B6 Distributed TrackFrom Systems Management to Service Management - Crossing the
Chasm
The market today is definitely in a state of change - commodity offerings
no longer compete effectively and so even from an IT perspective, we see the
swing towards differentiation of products and services driven top-down by
the business where the balance between cost of delivery and quality of service,
time to market and competitiveness, rules. Session 2C1 Service ManagementThe Top Down Drill Down Low Down - Getting the street-level view
while flying at 10000ft
Balancing breadth, depth and business value is a perpetual challenge for performance
engineering. How much information did we really need to understand to answer
tomorrow's questions by yesterday? Stepwise Refinement is a well-established
approach to addressing this challenge but in practise, moving from the 10000ft
view to the street-level has usually meant landing the whole aircraft and
shifting all the passengers to a different mode of transport. Session 2C2 Service ManagementWhere's Waldo: Uncovering Hard-to-Find Application Killers
We've all heard about performance anti-patterns, yet how do you find these
problems in your system? I have used Rational Quantify for over ten years
and have learned many tips and tricks to help uncover common anti-patterns.
This paper will describe the wealth of data generated by Quantify and how
to use this data to uncover software performance problems such as excessive
memory allocation, the ramp, unnecessary processing, one lane bridge, and
more. Session 2C3 Service ManagementPrediction of Performance Bottlenecks in Scaling Java Enterprise
Systems Software architects and managers need to know how the response time (or throughput) deteriorates as user workloads increase or the mix varies. This will allow reasonable claims about performance in both the design of new applications or enhancing or upgrading existing software. This paper will describe a technique to identify: 1 at what workload the software fails to properly scale; 2 which Java methods are the cause of the performance deterioration; and 3 how much optimization needs to be performed in order to bring the system back to its target response time. This will provide reassurance to the architect, or the manager, that the application will continue to scale and that all potential bottlenecks have been eliminated. Session 2C4 Service ManagementThe Holy Grail: Building Applications That Can Survive the Unpredictable
Web
We have all seen Web applications fail in spectacular ways under heavy load.
This paper examines approaches that can be used to understand and eliminate
this problem. Session 2C5 Management ITIL TrackHow's Your Memory? Detailed memory analysis of a system, especially a large complex system, can be very beneficial. The analysis may uncover performance issues or misuse of this limited system resource. Obtaining the analysis data though is often difficult. This paper describes how SAS uses RootCause. along with a custom probe to gather data from the SAS System memory management routines. The type of data collected will be described, along with the types of problems that can be uncovered by analyzing this data. Session 2C6 Management ITIL TrackPerformance Monitoring of Enterprise Java Systems Good practice for performance engineering recommend that there is monitoring of enterprise Java systems throughout the development cycle. This can prevent the sudden occurrence of major performance bottlenecks during production testing which often leads to lengthy and expensive project delays. Performance monitoring tools can be used to trace events that occur during run time and can attribute the cost of the events to specific parts of the application. This paper discusses a number of different monitoring techniques to collect performance metrics on large multi-user enterprise applications and shows how they can be applied to assess the performance of such systems through out the development cycle. Session 2D1 & 1D2Capacity Planning 101
What's It's About? Session 2D3 General TrackActive Baselining in Passive Environments In order to decide if systems are running according to their usual trend, it is necessary to compare against a performance baseline that defines the usual operating envelope. This paper describes how a baseline can be derived and actively updated from passive stores of performance data. The passive stores are typically flat text files or databases. From these, baselines can be derived as needed, varying the baseline norm, granularity, update frequency, etc. The outputs of baselining 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 the kinds of reports that can be produced and analyses that can be done. Session 2D4 General TrackHow to integrate capacity management into the application development
lifecycle This presentation will start by discussing the general principals that should be considered when planning to integrate capacity management processes into the application development lifecycle. It will identify the actions that should be undertaken and discuss various approaches that could be appropriate. It will then go on to describe how this is being implemented in a particular client environment, describing what interface points have been developed and what processes are performed at each stage in the development lifecycle. Lessons learnt during this implementation will be shared with the audience and there will be time allowed for a question & answer session to enable discussion of the subject area. Session 2D5 General TrackNo wolves, some rats and thousands of mice; a case study in
workload analysis Major system conversion projects have transformed Severn Trent Water's enterprise computer systems. However, legacy enterprise systems have continued to demand high levels of CPU. In order to meet corporate objectives for reducing business operating costs, a project to minimise CPU demand in the legacy environment, without adversely affecting live services, was required. This presentation tells the story of how CPU capacity was reduced without inflicting excessive pain on our users Session 2D6 General TrackCost-effective Capacity Modelling: A Real-life Case Study
This presentation explains how Excel was used to develop a detailed capacity
model for a key system WEDNESDAY 28th JUNE 2006Session 3A1 MainframeFlashCopy: a Picture that Takes Some Time to Develop The 'Flash' in FlashCopy suggests that this technology produces a mirror image of a volume in no time at all. And in a way it does: the volume and its image are both almost immediately available. But there may be a catch: write requests issued to either the source or the target volume may encounter 'below the surface' activity that may delay those write operations. This presentation will review how FlashCopy works, under what conditions and to what degree it may affect write performance and how this performance impact can be detected. Special consideration will be given to the use of FlashCopy before PPRC or XRC re-synchronization. Where possible the presentation will include charts generated from actual production data. Session 3A2 MainframeMining Performance Gold From CICS Statistics This session includes presentation of the essential CICS statistics for performance management and capacity planning activities. For maximum effectiveness on the job, attendees will learn (a) important considerations for parameters affecting the data collection, (b) the minimum set of reports required to support a particular activity, (c) what are the important fields on the key reports, and (d) how to avoid some potential pitfalls. Samples of the most useful reports will be presented. The emphasis will be on quick techniques that help us "mine" the mountain of information collected by CICS. Session 3A3 MainframeNetwork Performance & Availability Reporting: Someone Has to Start
It The measurement of network service levels has long been a neglected practice by the performance industry. This is most likely due to a lack of network skills training and standard measurement and reporting definitions, as well as the fast pace of emerging technologies. There are, however, some TCP/IP utilities that present a practical opportunity to measure network service level goals proactively. With some development effort, nominal coding and collaboration with other network tools, these TCP/IP utilities are very useful. This paper shows how such utilities may be used to gather network service level statistics. Sample reports are also provided. Session 3B1 Service ManagementWhere should I start my ITIL implementation? Awaiting Abstract. Session 3B2 Service ManagementManaged Infrastructure to Meet Your Business Objectives Will discuss some quick wins which could be achieved whilst trying to align your Storage Infrastructure to meet defined Business Objectives. A practical description of how one Utility Company re-aligned their IT internal processes in order to meet the newly stated Business Objectives. This will give an insight into the how's the why's and lateral thinking which helped achieve the goals. It will show how they reduced "time to market" and improved their quality of service and Service Level Agreements. Session 3B3 Service ManagementITIL Service Management - A Candidate for Business Process Outsourcing?
As the ITIL framework is increasingly adopted within the UK, and globally,
as the 'standard' for IT Service Management new implementation methods are
being considered which are different to those of the early adopters. Whilst
organisations have, since the mid to late 90's, considered outsourcing their
'helpdesk' function, most were reluctant to consider outsourcing other aspects
of their ITIL implementations. Session 3C1 End to End ManagementLies, Damn Lies and End to End Statistics
At Friends Provident we have made extensive use of End to End Response Time
statistics to help manage the performance of our key applications. This paper
details: Session 3C2 End to End ManagementEnd to End Performance Monitoring Awaiting abstract Session 3C3 End to End ManagementMeasurement of Transaction-Based End-To-End Response Time in Un-Armed
Environments 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 paper 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. Session 3D1 GeneralUtility Computing Awaiting abstract Session 3D2 GeneralAutomating IT Capacity Management The ever growing complexity and increasing size of current IT infrastructures are the top challenges for today's IT Capacity Managers. Facing several hundreds of vital business servers on a large variety of platforms makes Capacity Management automation a must. The presentation introduces a standard and straight-forward Capacity Management process and demonstrates how to implement the process steps. The presentation includes how to report on resource usage, deal with issues of scale, handle 'what-if'-scenarios, plan for future resource usage based on business drivers, and how to produce a standardized Capacity Plan. Session 3D3 GeneralCritical Success Factors for IT Service Delivery - Theory vs. Reality Despite of the resources invested in IT, many companies do not believe that they receive the service they hoped for. This presentation is based on my research into defining and understanding the factors that are vital to ensuring service delivery. The main focus of the presentation is the results and conclusions of an independent survey carried out by the presenter with 50 UK IT Managers and Directors in August 2005. The challenges identified include IT's involvement in strategic planning, control and complexity of the infrastructure, adoption of ITIL and the question of trust between the IT department and the business. |
Upcoming Events
2008 Sponsors
|
||||||||||||
|
©2008 UK Computer Measurement Group |
