Complex IT management, budgets, deliverables on time, utilisation, throughput, response, and availability; all issues we as IT professionals deal with on a regular basis. We know where each of these key areas are in terms of providing efficient effective IT services but unfortunately it is often not up to us to make final decisions concerning these matters. Executive management require a regular "feed" of information that they can use to manage their resources. This paper takes a look at establishing that "feed" of information to management.

How much should you invest in assuring that your business critical and customer facing applications will meet their service level metrics? A Performance Test Centre can provide predictive application performance analysis, but can you afford the investment?

This presentation describes the need, benefits and options for utilising a Performance Test Centre to assure the performance of your critical applications. Specifics that will be covered are:

• Background - Testing Today
• Why is there a need for a Performance Test Centre?
• What services does a Performance Test Centre provide?
• What are the Benefits?
• What are the Options?
• Option 1 - Do Nothing - accept the risk
• Option 2 - Build an internal Testing Centre - costly
• Option 3 - Use a purpose built facility - cost effective
• Key Issues
• Key Success Factors
• Summary

Successful organisations in this dynamic and evolving environment are re-thinking how business can achieve increased value from Information Technology (IT) by:

• Recognising IT systems are now visible to external customers and partners
• Recognising IT can enable new business initiatives
• Using business transaction value as a critical service metric
• Leveraging existing multi vendor software investments
• Optimising, integrating and automating IT service delivery mechanisms.

This paper examines why it is essential for organisations to achieve IT service excellence, and describes the challenges faced both in the construction and delivery of IT services evolving towards a business centric IT model.

It assesses the changing role of IT service delivery management, and highlights the impact of service delivery failure upon the organisation. The paper analyses why many organisations are ill-equipped to deliver IT service excellence and proposes a pragmatic approach through which organisations may manage this process to meet/exceed user/customer expectations.

Business today is volatile - characterised by mergers, takeovers, structural changes, new technology and outsourcing. In particular, too many large IT projects are technical successes but business failures. Often these failures are the result of insufficient attention to the people issues related to the management of change. This paper looks at the hidden costs of change, and some of the reasons why change is resisted. And offers suggestions for more effective management of change.

BS15000:2000 is the recent standard introduced for IT Service Management by the British Standards Institute. It is a practical, common sense, best practice specification for ITSM and supports the Information Technology Infrastructure Library (ITIL) process guidelines. However, its introduction and acceptance for an outsourced environment may be anathema to external service providers. This presentation will introduce BS15000:2000, cover both practical and hypothetical aspects of its adoption in an SLA, and address a customer perception of short-term gains and long-term aims.

BS15000:2000 includes the standard, Code of Practice, and Self-Assessment Workbook and encompasses ITIL principles and philosophy. It is a management summary of the processes and relationships between:

• Supplier Management, Customer Relationship Management,
• Service Support (Help Desk/Incident Control, Configuration Management, Change Management, Problem Management, Software Release Management)
• Service Delivery (Capacity Management, Availability Management, IT Service Continuity Planning, Security Management, Service Level Management/Service Reporting, and Financial Management).

IT Strategic planning without Measurement might also be called daydreaming. There are always cases where reality has been bypassed in the effort to come up with a suitably inspiring strategic plan, typically ending in a suitable disaster. Even at this early planning stage, measurement is required to give credence to any strategy being proposed. This ounce of reality is needed to get the serious buy in of business management as to what is being proposed, how long it will take to reach fruition, and approximately at what cost.

With the advent of corporate globalisation and 'e'-business, increasingly IT organisations are called on to demonstrate the value they deliver to the business. This invariably has a major impact in the way in which they operate as both a supplier and customer. In the past data centers were run as huge cost centers that at best, broke even or at worse, run at a loss. There were huge operational inefficiencies and inevitably bad customer experiences. This ultimately resulted in a large number of corporations outsourcing their IT organisations to specific companies specializing in IT delivery (facilities management). This is not so different in today's market place, especially in the area of e-business with Internet Service Providers (ISP's), Application Service Providers (ASP's) or Management Service Providers (MSP's).

Many of today's IT organisations outsource certain activities such as Storage Management, desktop environments or networks. But not all corporations want to outsource, in fact many are spinning off their IT organisations as a separate (but still wholly owned) business with their own strategies, goals, objectives, budgets, targets and management teams.

In these modern IT organisations, their customers and customer services are the key to success. Not surprisingly IT organisations are now concentrating more than ever on the quality of service and the cost of production.

IT customers on the other hand are more concerned with the purchase costs associated with IT and also the quality of their delivered service.

The IT framework that ensures customer satisfaction is a 'Service Level Management Framework' (SLMF). This paper highlights the major issues concerning the service level management dilemma of the modern IT organisation and how the SAS SLMF fits into the picture, based on our experiences at customer sites.

Project management in the Internet space has several unique challenges. Primary amongst them are the short time to market, critical nature of user interface design and usability, and ever changing requirements. Success is determined by the project team's ability to quantify, qualify and react effectively. The presenter will examine these issues and offer insights gained first hand as a project manager in this challenging environment.

Workload Manager is central to two new functions in z/OS: Intelligent Resource Direction and the IBM License Manager. These new capabilities are in some ways complementary and may also be in conflict with each other.

We'll examine both IRD and ILM in detail, describe the possible areas of interaction, and suggest ways to ensure harmonious operation of z/OS in LPAR clusters with Variable Workload License Charges.

The 1990’s was a period where outsourcing of IT services threatened to end the existence of all internal IS organisations - what did we learn from the Business Enterprise, IT Organisation and Service Provider perspectives:

• What lessons have we learned from all this?
• What are the key trends and issues for the future?
• What has been the government influence in outsourcing?

This session looks at the changes in the role of the internal IT organisation in the enterprise as well as the influence of the evolving outsourcing model from full scale outsourcing to selective outsourcing and the emerging importance of shared technical services in this industry trend.

Key topics covered in this session are:

1. The changing role of the internal IT organisation in the 90’s
2. The trends in outsourcing - what did we learn?
3. What is strategic sourcing?
4. What are the implications to the IT organisation and external service providers?
5. The role that shared technical services can play in an outsourcing model
6. Key success factors for establishing a balanced selective outsourcing model in the IS organisation

This paper provides insight into the Performance aspects and the Tuning possibilities for an OS/390, WebSphere, CICS/TS, MQSeries and DB2 Java application. The application uses Java totally for the front and back end. Application interaction is via a browser based Java applet to a WebSphere WAS servlet, to CICS via the CICS Transaction Gateway and then to DB2 via SQLJ and JDBC. Background processing uses MQSeries and triggered JCICS transactions. The talk discusses mainly DB2 V6 and CICS/TS V1.3 but will aim to include early experiences with DB2 V7, CICS/TS V2 and WebSphere V4.

Barry Merrill's MXG has been the cornerstone of capacity planning for many organisations. It collects data from many sources - not only MVS anymore.

So why run MXG on the mainframe?

With the increased costs of running SAS on the new technology what options are there?

This paper reviews the planning process behind moving and looks at issues and problems in actually making the move from one site's experiences.

This is a joint paper to be presented by David Vasey and Steve Jack of SJCC Pty.

MVS - From 9672 to zSeries

One site's experiences of moving from the CMOS 9672 hardware to the new zSeries that was released August 2000.

As sites around the world awaken to the dawning of IBM's zSeries, and given the decrease in hardware costs and increase in software costs, capacity analysts will require a better understanding of their mainframe environments prior to determining provisioning requirements.

This presentation will review the potential costs and challenges associated with the move to zSeries. It will present what options IBM and their partners have created to charge sites for their new hardware and software in the future and pays particular attention to the areas of:

• Workload License Manager
• zSeries Hardware

The presentation will also look at the challenges presented to the capacity planning/management teams within a site after the migration, in particular the additional costs when workload requirements are not determined correctly.

What is virtulisation and why is the term having such a high focus in the IT industry? Do all uses of the word virtualisation have the same meaning? What has virtualisation to do with RAID (if anything)? Where should virtualisation be implemented - in the SAN, the Server, or the Storage system? Why should anyone be interested in virtualisation and what can it do for people's businesses?

The first reference to virtual within IT that most people can remember was in the early 1970's when IBM described the capability to have logical access to more computer memory than physically existed as "virtual storage". Today's usage of the terms virtual and virtualisation has not fundamentally changed since the 1970's. Broadly speaking, virtualisation can be defined as the creation of an image or a "reality" that doesn't physically exist.

This paper looks at virtualisation in the 2000's by introducing a methodology described as the Layers of Virtualisation. This methodology identifies where virtualisation brings value to the areas of availability, capacity, performance, and storage management. The ultimate objective of this paper is to help identify answers to the questions posed above.

The OS/390 environment disk IO activity tuning opportunities can produce results that seem like miracles. Productive capacity of a complex can be greatly increased because IO activity tuning will also improve processor, main storage and virtual storage utilization. This presentation will demonstrate all that is needed for an effective IO activity tuning project along with the benefits achieved by such activities.

The presentation’s five major points are:

• Anatomy of IOs
• Where, What, How of Measurement Data
• Obtaining Information from Measurements
• Tuning Alternatives
• Sample Results

In less than 12 months, Suncorp Metway has grown from 1 TB of NT data in the SAN to over 5 TB. Of this, half is at the main site and half is at the DR site. The two sites are connected via dark fibre and WDM links. Remote Disk Mirroring is easy (sic). The difficult piece of the puzzle is how to put the DR servers back together with the copies of the production data at the DR site. This case study will examine the complexities of NT mirroring, stripe sets, recovering file servers, Exchange servers and SQL servers as well as other NT exotica.

I/O performance continues to play a central role in many applications, especially those involving very large databases, such as found in Data Warehouses. High performance arrays, ever larger disks, and fast interconnects enable new applications and services, but are only part of the myriad of factors that affect performance. The paper discusses the measurement and analysis of I/O performance, particularly in an open-systems environment. It covers both hardware aspects and the impact of the software layers typically encountered, such as device drivers, logical volume managers, file systems and application interfaces.

In an effort to minimise the total cost of ownership of enterprise computer systems an industry consortium (including Intel, IBM, Sun, Microsoft, BMC Software, and others) have sought to establish a complete set of management infrastructure standards that facilitate the integration of various hardware and software management systems. This initiative is called Web-Based Enterprise Management (WBEM) and is now coordinated by the Distributed Management Task Force (DMTF). In this paper we cover the overall goals and framework of WBEM, as well as critically assessing both its strengths and weaknesses. We then provide a more concrete focus and example by examining Windows Management Instrumentation (WMI) which is Microsoft's implementation of the WBEM standards.

History has shown that over 35% of web applications perform poorly due to issues that reside outside of an organisation's web infrastructure. This presentation outlines the benefits that your organisation can achieve from utilising hosted web load testing services to thoroughly test your web application's performance from outside your organisation's firewall.

A web-based, hosted load testing service conducts full-scale testing of your site in a matter of hours. By emulating the behavior of thousands of users against a staging server, it identifies bottlenecks and capacity constraints before your customer's site goes live.

This presentation discusses server consolidation which is becoming an area of increasing interest to organsiation suffering from server sprawl and tightening IT budgets.

This presentation looks at:

• What server consolidation is and why it is of interest.
• The various types of consolidation that could be considered.
• One approach to server consolidation.
• Some examples of some very large consolidation projects and what they achieved.

Tools for direct measurement of Quality of Service (QoS) as represented by End-To-End Response Time (ETE RT) are gaining acceptance. However, classifications of such tools were developed before the wide commercial adoption of the Web and are, thus, out of date. In this paper, author presents an update to Tsykin-Langshaw '98 classification, with the specific emphasis on the measurement of the Web-based activities. Review of the existing tools is included and specific difficulties of operation in the Web-based environment are discussed. Future trends are indicated.

A typical web application involves multiple processing tiers: browser, web server, application server(s) and possibly separate database server(s). Performance measurement in such an environment (often called n-tier) is sometimes not performed well because it is perceived to be a complex measurement problem. However, good data can assist in measuring performance against objectives during development; as part of a service level agreement after deployment; and during problem diagnosis if/when performance problems arise.

This seminar will use an open source n-tier architecture as an example and to explain the concepts involved. The same lessons will apply to most environments.

Several case studies will be reviewed in order to highlight the challenges and possible solutions to the application measurement question.

Participiants are encouraged to bring their own experiences for discussion. The seminar is aimed at the application and middleware levels, and will only make passing references to TCP/IP network performance.

This presentation will step the delegate through a complete Storage Area Network (SAN) design and implementation process. The workshop is based on a CASE STUDY in a large financial organisation trying to move from older technology UNIX systems (AIX and Solaris ) in a non-SAN (SSA / SCSI) environment to a scalable fibre based solution. The phases that will be examined include requirements, analysis, high level design, detailed design, tendering, evaluation, costing and finally selection. The focus will be on both technology considerations (physical infrastructure, control and management, technical issues) and financial implications (3 and 5 years TCO models compared to current). This workshop will give the delegate a process and issues to consider for both non-SAN and existing SAN environments.

Performance management controls of CICS for MVS/ESA (a.k.a. z/OS) and CICS Transaction Server (TS) for OS/390 regions can greatly affect the effective use of computer resources, and, as a result, the effective capacity of a CICS complex is impacted. This presentation will focus on the parameters and OS/390 environmental factors which affect a CICS region’s total demand for processor capacity, real storage, virtual storage and disk input/output service. Because of their significance, Workload Manager (WLM) Release 10 changes that impact CICS will be included. The potential risks and benefits associated with the selection of actual and default values will be identified.

Capacity and performance practitioners, particularly on mainframes, have used the SAS system for years to do their analysis and reporting. Microsoft Excel has been around for years and is widely available but its surprising how few people are exploiting its capabilities with regard to analysing and reporting on capacity and performance data.
This presentation gives practical advice on how to do a multitude of tasks using Excel on a PC. It is platform independent (various sources of data will be used in demonstrations) and focusses on practical advice. Topics covered include;

1. Pivot Tables: Manipulating data quickly and effectively; Dealing with gaps in data; Summarising; Creating interval records from event records

2. Filtering: Custom filters

3. Batch Modelling: Look-up tables; Gantt charts

4. Surface Plots: Combining 24 hour profiles with monthly growth

5. Forecasting Functions: FORECAST; GROWTH; Moving Averages; Linear Regression with Trendlines and the Data Analysis Wizard

6. Creating GIFs: Saving charts separately from the workbook; Creating HTML

7. Automation: Creating Visual Basic Modules through keyboard Macros

8. Charts: Dealing with ‘too much’ data; General tips

Getting to goal mode is no longer new news. People might be starting to panic in view of the withdrawal of support for compatibility mode in z/OS 1.3. Now, there is an inevitable reason to go to goal mode. You don't want to be last and without support! This Workshop captures the successes of others who have gone to goal mode and describes the migration process in mind-numbing detail. Topics include:

• What to expect from goal mode
• New and changed interfaces (operator, JCL, SDSF, RMF, ...)
• Goal mode vocabulary
• Approaches to conversion
• Pre-conversion preparation
• The IBM Goal Mode Migration Aid (GMMA)
• Detailed conversion steps
• Anatomy of a Service Definition
• Getting CICS to play nicely with others
• Advanced goal mode features