AGENDA

Digital Data Directions

Fred Moore
Horison Information Strategies, Inc

However you define it, in the 50 years since the development of the first computer, the capture, manipulation, transmission and consumption of information in digital form have become critical functions in our economy and soon, perhaps, in our civilization. For years, the spectacular proliferation of digital computing and networking has been rewriting the rules of business, and it will continue to do so. There has been more improvement in the fundamental elements of digital storage since 1990 than there had been in the history of the data storage industry. The rate of change has been unprecedented. The mainframe that was supposed to be dead in 1990 has become scalable, less expensive and easier to maintain. This growth of the more traditional applications is bolstered today by an accelerating move to convert video, audio, pictures, medical images, paper copy, visualization and plain archival data to digital storage. We are now confronted with terabyte and soon petabyte storage applications to manage multiple compute platforms that may be geographically distributed. Traditional challenges that the mainframe market successfully addressed in the past are now challenges for the distributed clients; backup, recovery, performance, availability, and security are the major challenges for the rest of the compute platforms to resolve. These changes have now signaled that the march toward actually sharing data among dissimilar computer platforms remains the No. 1 unmet challenge for the computer industry to address in terms of customer requirements.

This paper will highlight some of the key issues related to digital storage as well as the underlying technologies that are enabling unprecedented growth. The Information Revolution promises to touch and possibly transform many aspects of the life we know through books, newspapers, phone calls, television, movies, musical recordings, and architectural drawings and perhaps in ways we have not yet imagined.

Understanding Change Management in a Data Centre

Pierre LOUYS
State Rail IT

[Prevention is better than cure...]Today, for a reliable organisation, simply fixing what has gone wrong in the production line or the delivery of a service is not enough. The management and customers' requirements demand effective actions to prevent error occurrence and recurrence. Change Management in IT Many IT providers struggle to address these requirements and spend as much as 25% of their resources to find and fix problems, deploying automated solutions in order to : · control the integrity of their IT infrastructure · manage their complex and distributed technologies. Changes affecting the integrity of the IT infrastructure can be broadly divided in three categories:

1. changes introduced before a "system" becomes operational
2. changes implemented after a "system" is operational
3. unplanned changes (due to equipment failure, human error, disaster).

The first category is a matter for technology suppliers and application developers. The third category is a problem management issue. This paper is only concerned with the second category and it explains how it can be addressed in a Data Centre environment.

Capacity Management a Methodology

Adrian Heald
Capacity Reporting Services

Primitive Capacity Planning measures the "fullness" of each IT processing Component and recommends action before they become "full". This ensures that there will be sufficient resource to carry out business functions but can lead to upgrading processing components that that may not need upgrading or are not the most cost effective to upgrade. More sophisticated methodologies model user response and recommend action before these are impacted. This paper presents phase 1 of a 4 phase methodology that provides an orderly progression from "no capacity planning" through primitive "fullness" capacity plans to sophisticated response based capacity planning.

How to get the most out of your AS/400

David Pickett
Pickett Computer Services Pty Ltd

Many large organisations have AS/400s for special purposes but don't know how to properly measure and improve performance. This presentation will give an overview of tools available to assist in proper performance management, including no-charge tools that are a standard part of OS/400.

Why S/390

Phil Smith
IBM Australia

The marketplace is changing, if it has not already changed. Customers now basically want to be able to connect to any application, at anytime and anywhere. This may sound like a fundamentally sound process to adopt, but it puts pressure back on the organisation supplying the service. Choice of Operating System, Data Base, security package, hardware etc all adds confusion and doubt to the decision process. So why should any one platform stand out more so that any other when running "Core Business" applications.

Revamping Your Capacity Management System

Adrian Heald
Capacity Reporting Services

Capacity management is mainly about providing the decision makers of an organisation with information they require to make those decisions. Several processes must occur before the data collected at the monitored platform can be presented as useful information; data collection; data storage; data summarisation; data archival; and report generation. These processes are dependent upon one another (data collection before report generation etc.) and thus many establishments have them rigorously structured. This often causes problems when components fail, the most significant of these potential problems is data loss. This paper looks at an alternate method of structuring your capacity management system to eliminate the dependency between the various processes involved.

Implementing Software Asset Management

John Mycroft
Mycroft Systems Ltd

The process of putting Software Asset Management in place:

• Setting priorities
• Who to get involved
• Recording progress
• Publicising results
• and so on

End-To-End Response Time and beyond: direct measurement of Service Levels

Mike Tsykin
FBA Computer Technology Services

End-To-End Response Time (ETE RT) is the preferred metric for the measurement of service Level. It shows whether a User was prevented from working at full capacity. ETE RT is an indirect metric due to difficulties in identifying a transaction in Open Systems. Other measurements may be better suited to the task. One of many is the time when a User was capable of working but was prevented from doing so.

This paper reviews existing approaches to measurement of ETE RT, lists available tools, outlines the concept of direct measurement of Service Levels and describes an approach to its implementation

PSLC IN AN HOUR - CONSTRUCTING A SIMPLE PARALLEL SYSPLEX

Mark Heers
Amdahl

IBM have recently imposed more stringent requirements to continue to receive the benefits of IBM's more cost effective Parallel Sysplex Licence Charges (PSLC). In effect, this means that sites with multiple processors will have to implement some form of Parallel Sysplex sharing in the near future to continue receiving the benefit of PSLC. This paper explores the construction of a simple Parallel Sysplex. Whilst the paper does not recommend quick and unplanned implementations, it is designed to give an idea of the requirements of constructing a Parallel Sysplex. As in all construction, it commences with the foundations of a sysplex - the definitions of the signalling paths between the MVS images, the implementation of a common timer reference, the control of resource serialisation (under GRS) and the creation of a sysplex couple dataset. Upon these foundations, a set of guidelines and actions called policies are defined to describe the usage and resourcing of the Parallel Sysplex by various applications. Finally the paper describes the implementation of a sample application, namely the sysplex-wide recording of hardware and software failures (logrec error data).

"A house is a machine for living in" Le Corbusier (1887-1965), "Towards an Architecture"

INFORMATION NEEDS -V- PROTECTION OF THE INDIVIDUAL

Bruce Edwards
Data Security Services

The paper examines the needs of the organisation in retaining information and the corporate view of security balanced against the practical storage and access problems of IT. The issue of privacy, as perceived by us as individuals, and the view of confidentiality as seen by organisations is discussed leading to the legal requirements of record keeping as it applies to IT practice. The need to have a method of classifying data to determine which records are vital and how this relates to hierarchical storage systems is reviewed. The paper concludes with review of a recent European court case affecting records storage and presents the concept of "Virtual Records" in E-mail and Data Warehousing.

Client Server in a Government Agency! Five years of Failure

Bill Stewart

We wish to know whether a project is successful; but although everyone hopes to learn by a mistake, nobody likes to admit they have made one. One sort of project is to devise a strategy; and strategists avoid admitting they have failed by arguing, wrongly, that they have no responsibility for implementation. The mistakes they make include failing to select and promote management and technical methodologies, and to establish in sufficient detail a logical connection between the methodologies and the perceived and implicit requirements of the organisation. We consider the case of client server information technology in a government agency over the past five years.

Helping Your Business Help You With Capacity Planning

Bloomer for John Knight
Coles Myer Ltd

History:

• multiple spreadsheets of resources used, updated via cut and paste
• having to re-range graphs in Excel
• businesses asked few queries, often repetitious ones
• businesses not aware of resource usage of applications
• no automation

Sound familiar?

Well, we set about designing a new system that would help the businesses and ourselves based on Microsoft's Access. This meant cooperating with the businesses and trying to instigate what they really wanted. I found that I could predict this reasonably well by putting on my business hat. What does a business want to see the first time they start the application? They want to see how they are tracking against budget! We provided this facility and much, much more. Businesses that have this application (it has yet to be rolled out to all) are extremely happy with it and the others are eager to get it as soon as their software versions support ours. Having this tool has prompted real investigation into what causes resource budget blow-outs. Consultants have been called in to "Strobe" troublesome applications. Information is accessible by business applications people in a timely manner. We provided a package solution.

Client Server Volume Testing

Peter Taylor
ATO

Over the last year the ATO has been developing new applications using the Cool-Gen (ex-COMPOSER) RAD tool. These applications still rely on the MVS environment for CICS and DB2 but have given the Business people a look and feel of a window application. The challenge has been to maintain consistent support in the area of Volume Testing for these applications.

This paper will describe the joys and experiences of converting to Mainframe Client/Server testing and what to look out for when you bite that bullet. The audience will gain a better insight into the "How-To’s" of testing a Mainframe Client/Server application.

New Methods for Measuring Function Points in Outsourcing Contracts

Pam Morris
Total Metrics

Function Point Analysis (FPA) is used by organisations worldwide as one of the measures used to establish the baseline size of their software assets. This paper introduces new techniques, which enable all the functionality delivered and worked on by the supplier to be included in the productivity performance monitoring of these contracts. Typically only the business applications layer can be measured using FPA. The infrastructure software eg. Utilities, device drivers and gateway applications, are usually overlooked because FPA is not designed to, nor easily adapted to, measuring internal layers of functions not delivered to the business user. This new Full Function Point Technique developed by the University of Quebec in Montreal is a refinement of the FPA technique. It is no longer limited to only measuring MIS type applications but was specifically designed to meet the needs of organisations who build and support infrastructure applications and real-time and embedded software.

Vitual Tape Subsystems - Alive and Kicking in Australia

Ron Fellows
IBM Australia

VTS's have been running production workloads in Australian sites since April, 1997. This presentation will cover the practical experiences of several of these installations including implementation planning, data migration strategies, performance management, and disaster recovery issues. Come and hear the "REAL" story about this exciting new tape technology and like virtual storage and virtual disk before it how it is changing the whole world of S/390 tape processing.

The Only Good I/O Is a Dead I/O!

John Tyrrell
IBM

How to reduce that inveterate batch window? There are many functions and techniques in S/390 platforms that can truly help you greatly reduce that elusive (and growing) batch window, but how do you find those opportunities and quantify them? This paper will describe actual cases where the batch window was analyzed,recommendations made and implemented, and then later re-measured to verify the study analytical results. In one case, a 6.5 TB shop was able to reduce their batch window by over 4.5 hours per night! This represented an overall 50% reduction in their critical batch window.

Trends in IT Outsourcing - Myths Exploded

Neil Fraser
Deloit's

The outsourcing of information technology services, particularly in the mainframe and enterprise server end of the market, is now a major and mature tool of management. This session will use the results of Deloitte's recent survey of IT outsourcing in Australia to look at why IT departments outsource, the benefits expected to be received, and most importantly the actual experience of organisations that have outsourced, and the lessons they have learned. The session will also look at how a structured approach to IT outsourcing can ensure that the expected benefits are realised, and that the many risks associated with IT outsourcing are minimised.

Network application Performance testing : Lotus Notes

Andrew Williams
IBM GSA Test Centre

As customers move toward network applications, the problems of acceptable performance and user load stresses have not evaporated. Increasingly, customers need to simulate large network user loads to measure end-to-end response time and identify potential bottlenecks. This paper presents a methodology in use by IBM Global Services Australia Test Centre to achieve this and details results from real applications evaluations on an AIX Lotus Notes 4.x environment. In all, thirty client machines were setup in a laboratory and linked to an additional set of seventy virtual users all exercising Lotus Notes application test cases to create a user load of 100 active concurrent users. During the process the clients and servers were monitored with various tools. The paper details the process used, a sample of the results, problems found in the process, the metrics required and future directions for network performance test solutions. It will focus mostly on the method of creating large, realistic loads for Lotus Notes applications - a topic largely ignored by the testing community up to this point in time. The subject will be of interest to any customer interested in implementing a network application or client/server application or a test specialist involved in testing these type of applications.

VIRTUALISATION OF THE WORLD

Gary Barker
IBM Australia

Virtual disk and virtual tape subsystems are being made available by several vendors in the storage marketplace today. They offer benefits to computer installations, in terms of service delivery and cost reduction and also in terms of functionality to the business. Their successful exploitation requires a mindset change; they need to be thought about differently, particularly from the point of view of performance management.

The concept of virtual storage is not new; processor virtual storage has become so fundamental to everything we do that the advantages and methodologies for managing it are now second nature. This paper first reviews the advantages and some of the performance considerations virtual storage brought to the processor world. It then explores the similarities between these factors and the factors that relate to virtual disk and virtual tape devices