This is a summary of a poster session presented at the 1993 CAUSE Annual Conference held in San Diego, California, December 7-10. It is the intellectual property of the author(s) and may not be published without permission. Disseminating the material otherwise is permitted, provided that the author(s) and the original presentation of the 
paper at CAUSE93 are acknowledged. 


   CONSIDERATIONS IN MOVING APPLICATIONS OFF THE MAINFRAME


INTRODUCTION

The situation at the University of Virginia is the familiar 
one of a mainframe being heavily utilized and the need to 
find ways to relieve this.  As Information Technology and 
Communication considered ways to alleviate this situation the 
proposal was made to have administrative end users who create 
ad hoc reports do their work on an RS6000.  My poster 
summarized an analysis of this proposal.

The preliminary research was to interview each of the fifteen
administrative departments having staff who write programs. 
These programmers include highly skilled people able to 
respond quickly to departmental requests.  Although ITC 
standards prevent them from writing to production data, their 
ability to produce reports is considered mission critical by 
their management.

Subsequent research was to examine data collected by the 
System Management Files for a typical week.  Naming standards 
already in place at ITC were used to determine whether data 
files had been created by production systems or by end users 
and which computer jobs had been submitted by end users.
 
The challenge in this analysis was in deciding what questions 
to ask and how to answer them based on something measurable.  
For the poster at CAUSE93 the questions were stated, answered 
in summarized form, and by a table or chart.

CONSIDERATIONS 

Who are the users and how active are they?  This was 
accompanied by a table showing information on fifteen 
different departments.  The summarized data revealed constant 
activity, frequent use of production data, and numerous 
tape/cartridge mounts.

Are the end users consuming enough resources to make it worth 
moving?  This chart showed so much activity, especially 
during the normal working day, that it made further study 
worthwhile. 

How is the data needed by the application being stored?  A 
chart showing the number of accesses to production files, 
user files and tape mounts followed.  The large number of 
accesses to many different data sets showed that providing 
access to these files from an RS6000 would involve 
transferring and duplicating large amounts of data.
 
Will data be shared with applications remaining on the 
mainframe?  In this case, yes.  A chart just showing accesses 
of production data confirmed that the end users needed 
constant access to production files that would not be moved.  
Data would need to be duplicated on the RS6000.

Is the data being stored in a format that can be used on an
RS6000?  Certain file types prevalent on the mainframe cannot 
be used on the smaller platforms.  In order for these to be 
made available to the end users the data would need to be 
rearranged.  Further it involved numerous data files.

CONCLUSION

The analysis showed that the administrative end users needed
constant access to large amounts of production data.  Despite 
the significant amount of activity generated by the end 
users, this dependence on the data and the technical 
considerations in making it available on an RS6000 resulted 
in this project being dropped.  My recommendation is that as 
production systems need to be replaced platforms other than 
the mainframe be considered.  Then the end users needing data 
from that system follow.

Cause Comments

This was my first CAUSE conference and the first time I 
presented a poster.  The experience was wonderful.  

Margaret Massey of Miami Dade Community College deserves 
special credit for the organizing the poster sessions.  She 
combined creative planning with warmth and friendliness.
 
Particularly helpful to me:

     * Prompt answers to all my questions during preparation.
     * The special breakfast on the morning of the poster to
       meet the other presenters.
     * Having afternoon refreshments in the same room as the     
       posters, insuring a large attendance.
 
Suggestions to people about to bring posters, based on my
observations and conversations:
     * Request more than one 30 x 40 inch board if necessary.
     * Have a handout briefly describing the poster.
     * Assemble and laminate the poster before the conference
       (these were carried in artists' portfolios)
     * Allow a little more time to set up for people not 
       bringing already assembled posters.


Tracy Scharer  
User Support - Information Technology and Communications    
University of Virginia
804-924-4188
scharer@virginia.edu


Date Revised: 02/24/94 (bdz)