Spring '98 FORUM
March 5-6, 1998
Sheraton Denver West - 360 Union Boulevard, Lakewood CO (303) 987-2000

AGENDA THURSDAY MARCH 5 - 
8:00-8:30	Registration

8:30-9:00	Annual Business Meeting

9:00-10:00	"Analysis of a Problem in Goal Mode Scheduling"
		Ethan Bolker, UMass Boston and BGS Systems 
		Dr. Jeffrey Buzen, BGS Systems 

This talk presents some initial results from our analysis of Goal Mode scheduling 
under Workload Manager in MVS Version 5 and OS/390.  Goal Mode scheduling can be 
regarded as an additional level of control that operates above a conventional 
priority scheduler.  Instead of specifying priorities for different classes of 
work, administrators specify performance goals they would like each class of work 
to attain.  We discuss how to determine whether a set of response time goals is 
feasible, and, when it is, how to find a set of priority assignments that enable 
those goals to be met.  This talk focuses on the insight our analysis provides 
rather than on the mathematical details of our solution.

10:00-10:15	Break

10:15-11:15	"Capacity Planning and Performance Issues in Y2K Projects"
		Ellen M. Friedman, SRM Associates, Ltd.
		Jerry L. Rosenberg, SRM Associates, Ltd.

Almost all major IS Departments are faced with monumental efforts in preparing for 
required Year 2000 system changes.  Unfortunately many are not aware or are not 
planning for the potential capacity, performance, and financial implications of those 
changes.  This paper will explore the background of the Y2K problem, discuss the 
management structure that should be in place, and detail the exposure that potentially 
exists.


11:15-11:30	Break

11:30-12:30	"Year 2000 Testing Experiences, or 666 Days Until Nothing Works"
		 Michael L. Puldy, IBM Business Recovery Services Y2K Delivery

This presentation will discuss Year 2000 testing experiences.  No platform is exempt.  
Large Systems, Mid-Range and Personal Computers are all impacted by the computer event 
of the millennium.  Some companies make the transformation themselves while other 
companies contract out to vendors to make the magic happen for them.  Successes and 
failures will be described as well as some testing and assessment strategies that may be 
right for your company.  Financial implications and staffing complexities will also be 
reviewed.

12:30-1:45	Lunch - Sponsored by Candle Corporation

1:45-2:00	" A Proposal for an RMCMG Working Group"
                Dr. Tim R. Norton, MCI Corporation

This presentation is a brief description of the proposed RMCMG Working Group on the 
Coordination of Industry and Education for Computer Measurement.  This working group will 
establish an interactive relationship between Rocky Mountain companies and 
colleges/universities to promote computer measurement activities.  The overall goal of the 
working group is to encourage the universities to offer the needed courses and business to 
provide research opportunities in current problem areas.

2:00-3:00	"Computer Printing Economics"
		Frank Bereznay, Automobile Club of Southern California

The economics of high volume production printing are very different from other areas in the 
Data Center; variable costs are far higher than fixed costs.  More importantly, the costs of 
printing itself are dwarfed by the costs of finishing and postage for output destined for the 
mail room.  The total cost of this operation needs to be analyzed to direct attention to the 
most important issues and opportunities.  We provide a summary of the total costs for printing, 
finishing, and mailing, and we show how the information can be used to identify the real issues 
to manage the operation better.

3:00-3:15	Break

3:15-4:15	"Application Modeling Methodology"
		John Pilch, Performance Capacity Solutions

All planning exercises, at their essence, consist of answering four fundamental questions: 
1) Where are we?  2) Where are we going?  3) How do we get there?  4) What  will it cost?  
These fundamental questions are answered when the focus is capacity planning and configuration 
sizing.  This presentation will show a detailed approach to determining the resource requirements 
for new applications including capacity planning and evaluation of the application design.

4:15-5:15	Cocktail Hour -  Sponsored by Storage Technology Corporation



AGENDA FRIDAY MARCH 6 - 

8:30-9:00       Registration

9:00-10:00      "Understanding OS/390 Coupling Facility Measurements"
		Peter Enrico, IBM
		
The objective of this presentation is to enable the attendee to feel comfortable reading 
and analyzing the RMF Coupling Facility activity report.  The presentation will introduce
the attendees to the Coupling Facility measurements available by RMF Monitor I, explain 
concepts like SYNC and ASYNC, and then shed some light on Coupling Facility performance 
analysis in an OS/390 Parallel Sysplex environment by reading examples of the RMF Coupling
Facility activity report.
		 
10:00-10:15     Break
		 
10:15-11:45     "A Collection of Diverse Technical Notes"
                 H. W. "Barry" Merrill, PhD, Merrill Consultants
				 
1) Measurement of CPU Utilization in PR/SM, MDF, and MLPF.   2) Using Reporting Classes 
(or RPGN's) for workload measurement.  3) Merging CICS and DB2 by UOWID for transaction 
accounting.  4) SMF Interval Start Time is earlier than TOD by leap seconds.  5) What 
happens to measurements in a Y2K Test system.
				  
11:45-1:00      Lunch - Sponsored by Candle Corporation

1:00-2:00       "Windows NT Measurement and Tuning"
                Mark Friedman, Demand Technology 
	
This presentation is a review of the performance metrics available for Windows NT 
performance management and capacity planning.  It will also include a discussion of how 
key metrics can be used to resolve traditional performance issues.
								    
2:00-2:15       Break

2:15-3:15       "Workload Management and Open Systems"
	        Brian Wong, SUN Microsystems
	
One of the things that has always separated "mainframes" from "open systems" has been the
former's ability to run multiple workloads.  Open systems in the data center, whether 
running UNIX or Windows NT, are nearly always asked to handle a single workload, or even 
just a component of a workload.  From a technical standpoint, one of the keys to running 
multiple workloads is the ability of the MVS mainframe to allocate and manage resources 
for each requestor. Because the cost of managing and maintaining a data center is roughly
proportional to the number of systems, rather than their complexity, MIS organizations 
are now seeking ways to run additional workloads on open systems - or to transfer work 
back tomainframes.  This presentation considers the technical issues and reports on current
research in this area.     
										    



At press time, these speakers are confirmed.  RMCMG reserves the right to substitute without 
notice in case of cancellation beyond our control.
										    



Dr. Barry and Judith Merrill have again extended their generous offer to all RMCMG registrants 
and their families to join them Saturday, March 7, for a day of skiing at Keystone - they
will provide lift tickets at their expense.  Then join them apres ski at their condominium for 
light refreshments.  Details will be provided at the conference registration desk.