MEMORANDUM
April 30, 1996
To: Dr. David Schrader, Chairman of the Information Resources
Council
From: Bill Buntain, Chairman of the Ad Hoc Committee on
Distributed Computing Support
Re: The Ad Hoc Committee's Final Report
With the attached report, the Ad Hoc Committee on Distributed
Computing Support completes its charge. This final report does
not replace, but rather supplements the committee's interim
report. The appendix of this report contains some additional
notes on the interim report recommendations based on the
questions raised at the March meeting of Information Resources
Council. It should also be noted that the support levels proposed
in the Interim Report do not take into account future growth in
the number of machines deployed or users in the various segments
of the campus community.
Per the recommendation approved at that meeting, the Distributed
Computing Support Management Group should now be constituted. It
is our hope that these recommendations would be passed on to them
for them to further develop and oversee their implementation. Per
the comments in the IRC meeting, the new group should include a
representative of the Microcomputer Maintenance Shop. The same is
true of the Library. Their omission was an oversight.
I would recommend to the new group that they subdivide the
tasks associated with these recommendations and parse them out to
teams of central and distributed support staff to work on. For
example, there should be a team charged with reviewing and making
recommendations on an appropriate platform for a campus-wide help
desk application, a 32-bit Intel operating system and desktop
management tool, the set of applications to be included in the
standard setup, etc.
The effectiveness of the Distributed Computing Support Management
Group and its progress in accomplishing the recommendations put
forth by the Ad Hoc Committee should be reviewed by an
appropriate group identified by the IRC after one year. The
findings of that review should be forwarded to the Steering
Committee for evaluation of the success of the current management
model.
I wish to express my appreciation for the participation of the
members of the Ad Hoc Committee. I believe they have done an
exemplary job of cooperating in developing a far-reaching set of
substantive recommendations in a compressed time frame. I hope
that their recommendations will be put to good use.
Ad Hoc Committee
on Distributed Support Issues
Final Report
May 20, 1996
Executive Summary
In addition to the recommendations contained in its interim
report, the Ad Hoc Committee recommends the following steps to
leverage the existing and proposed distributed computing support
resources at the University:
Standardizing distributed server and desktop computing hardware
and software by:
Building a software distribution scheme around common network and
application software and a standard 32-bit desktop operating
system and acquiring one or more appropriate network management
tools, licensed on an enterprise level, to manage both desktops
and network servers;
Establishing a plan for a three year rolling window for
upgrading/phasing out desktop computing systems; and
Developing appropriate NetWare Directory Service and file server
standards.
Improving coordination of the efforts of central and distributed
support units by formalizing a two-tier support model with:
a central help desk function to support common applications and
coordinate support requests; and
a common problem/service request tracking system to be used by
both central and distributed units.
The recommendations contained in this report support two
primary objectives:
1. Increased responsiveness to user demands for support.
2. Leveraging of limited manpower support resources.
These two objectives are complementary. If manpower resources can
be effectively leveraged, the result will be an increased
availability for user support.
The strategies identified to accomplish these objectives fall
into two basic categories:
1. Standardizing server and desktop hardware and software.
2. Improving coordination of the efforts of central and
distributed support units.
Standardization
Findings:
1. There is considerable unnecessarily redundant effort
expended in:
Evaluating network and application solutions, including software
compatibility
Installing and configuring common network and application
solutions (such as word processors, spreadsheets, desktop
drivers, etc.).
Note: This finding does not diminish the need in distributed
areas for specialized applications that support their missions.
2. The diversity of desktop configurations complicates remote
user support (i.e., answering user questions without an on site
visit), desktop management, and migration to new application
platforms. A prime example of the latter phenomenon is the
complexity of migrating the Computing Center and Administration
Buildings to Virtual Loadable Modules (VLMs), NetWare 4.1 and
GroupWise compared with the relative ease of the same task
implemented in the College of Business Administration, which has
implemented a standard desktop configuration that is used by over
90% of its faculty and staff users.
Central support staff called in to look at distributed servers
have a steep learning curve, because every academic and
administrative area sets up their server differently. The same
problem faces distributed support staff when they attempt to
assist users.
Proposition:
Standardization of server and desktop hardware and software
configurations along with appropriate management tools will
significantly leverage distributed computing support resources
while at the same time increasing the quality of support those
resources provide.
Corollary:
The University must continue to permit diversity to support
the unique computing requirements of individuals and units, but
the cost of that additional support must be recognized and taken
into account in the budgeting process of that unit.
Recommendations:
1. The Computing Center should set up a server that will
act as a repository for pre-installed and configured standard
application and utility software installations under the
direction of the Distributed Computing Support Management Group
(DCSMG). The server drive(s) thus configured could then be
replicated to other servers throughout campus.
2. The DCSMG should identify a standard 32-bit
operating system platform as a strategic direction for migrating
Intel-based microcomputer desktop platforms and develop a
standard desktop configuration/administrative model to support
that platform. The Computing Center should play a leadership role
in this effort. The standard desktop configuration should
include:
network and video drivers;
a robust TCP/IP stack;
a LAN inventory client;
remote control software; and
files needed to enable applications (.INI files, etc.)
Distributed and central support organizations will be
empowered to give priority to troubleshooting problems with the
standard configuration on the premise that they affect more users
than customized setups.
3. The DCSMG should evaluate comprehensive network
management tools such as Novell's ManageWise and Microsoft's SMS
with particular emphasis on desktop management, including
configuration, inventory, and problem diagnosis. The Computing
Center should pursue enterprise licensing (either a site license
or a volume purchase arrangement) for the platform selected.
4. The DCSMG should look at establishing NetWare
Directory Services (NDS) standards in the following areas:
object naming and definition and
tree structure, including hierarchical design,
partitioning and replication.
These standards should be co-developed with the University's
X.500 standards.
Furthermore, the following steps should be taken relative to NDS
deployment at UNT:
Use of NDS directory map objects should be made a standard
practice.
A plan for migration away from applications dependent upon bindery emulation should be developed.
The central support organization should be given browse rights to all objects on the tree for purposes of backing up the tree and performing troubleshooting tasks.
A policy should be adopted that no NDS or X.500 schema
modifications should be made without the approval of the DCSMG.
5. The DCSMG should set NetWare file servers standards
for the following:
drive names;
drive mappings;
print queue setups;
directory structures; and
UPS and monitoring software.
Each network manager should be required to document their server
data backup and restore plan and execution of regular backups and
test restores must be part of that network manager's Planning
Guide.
6. The DCSMG should develop a plan for a three year rolling
window for upgrading/phasing out desktop computing platforms.
This plan should empower the Microcomputer Maintenance Shop to
leverage the University's purchasing power through discounts
available through placing bulk purchase orders. In the short term
particular attention should be paid to phasing out
support-intensive components such as problematic or obsolete
network interface cards.
Coordination
Findings:
1. Both users and distributed support units are confused
as to who to contact with particular needs/problems. In many
cases, the basic cause of a problem is not clear. For example,
being unable to connect to the administrative mainframe may be a
microcomputer hardware or software problem, a data communications
problem, a problem with the NetWare for SAA gateway, or a problem
on the mainframe.
2. Users and distributed support units often must coordinate
contacts with several units to carry out basic functions such as
getting a microcomputer connected to the campus-wide network.
3. Because of the uneven distribution of first-line
support resources in the departments and colleges, these units
receive varying levels of support from the central organization.
4. Despite the recent sprouting of World Wide Web-accessible
resources, there is not a standard and unified method for
distributing information to users and network support people.
Proposition:
Users should only have to make one call to get their
computer-related problems addressed and should be able to expect
a timely response.
Corollary:
The coordination of the support efforts among
distributed and central support units should be transparent as
much as possible to users.
Recommendations:
1. The relationship between the distributed and central
support staffs should be formalized along the lines of a two-tier
support model.
The central organization should be charged with:
long-range, campus-wide planning;
support of shared LAN resources (faculty/staff mail servers, mainframe gateways, the data communications infrastructure);
high-level technical support;
common application support and training.
The distributed centers should be responsible for:
on site user support;
daily management and operation of distributed servers;
maintenance of NDS and X.500 directory resources for
their respective units.
2. The DCSMG should oversee deployment of a comprehensive system
to incorporate help desk documentation as well as problem and
service request tracking. This system should be able to
coordinate the support activities of central and distributed
units involved in support of computing and communications and
should include:
universally accessible documentation of campus computing and
communications resources;
a list of currently supported revision levels of software, including applications, system utilities, NetWare Loadable Modules, and network and video drivers;
a referral list that identifies support responsibilities;
troubleshooting charts for common problems; and
a network connection request coordination.
Distributed areas should be able to use this system for tracking
their own support requests, as well as routing appropriate
support requests to the central support groups. The functionality
of this system should include:
a World Wide Web interface for users and distributed support
personnel to log calls and inquire on their status;
the ability to interface with an SNMP-based network management system for trapping communication network faults;
call prioritization;
call routing to individuals or to central or distributed support groups;
notification of calls via e-mail and a pager gateway;
automated call status tracking/escalation; and
a searchable call history.
These recommendations should not be viewed as
comprehensive, but are intended as a starting point for the DCSMG
to use in its evaluation of support system alternatives.
3. The central help desk function should provide:
application support for the applications included in the standard
configuration propagated from the executable servers; and
logging of all computing-related problems and service
requests with referral to the appropriate support organization.
Full-time staff roles and levels should be defined within the
help desk function to support:
data communications systems, including remote access;
UNIX systems;
messaging systems;
microcomputer desktop operating systems and applications;
the administrative mainframe, including gateway access.
Particular attention should be paid to extending the
hours support is available. (See Recommendation 4 in the Ad
Committee's Interim Report).
Note: The above two recommendations should not be interpreted as
proposing that distributed computing support organizations cease
to support such functions. Instead they are intended to
facilitate the coordination of the customer support function
between the central and distributed support organizations and
insure that a user can get assistance no matter who they contact.
It is anticipated that the distributed support organizations will
continue to take the lead in direct customer support functions.
APPENDIX
This appendix addresses several issues that were raised
at the March meeting of the Information Resources Council (IRC),
when the Interim Report was presented.
Classifications
In the March meeting of the IRC, a question was raised
as to how the various classifications for the proposed new
positions were determined. The classifications recommended for
full-time positions by the Ad Hoc Committee were based on the Job
Classification descriptions developed for Computer Support
Specialist positions through the Human Resources Office, as
follows:
Computer Support Specialist IV positions were recommended where the position had primary responsibility for support of a large and diverse network;
Computer Support Specialist II positions were
recommended where the position has secondary support
responsibility, i.e., the person in the position would be
reporting to a higher classified distributed support person who
had direct, primary responsibility for support of a network and
its users.
The rates for hourly positions were derived as follows:
$10.00/hour for a position doing both user support and custom
programming as primary job functions.
$8.00/hour for a position doing only user support.
As noted in the meeting, if the Ad Hoc Committee's
recommendations are adopted, existing positions should be
reviewed for pay equity issues.
Number of users/machines in distributed areas
The Committee was asked to review the numbers on which
the recommended allocation of staff positions contained in its
Interim Report was based. It was noted in the IRC meeting that
some adjustments in the actual positions recommended might be in
order depending upon the level of funding that was approved in
the budget process. The Committee waited for some indication of
the level of funding that would be available, but ultimately did
not participate in re-evaluation of the distribution recommended.
The issue was looked at by the deans, the Provost's Office, the
Senior Director of Academic Computing, and the Director of
Network and Microcomputer Services.
Comparison with positions requested in the Budget Process
The IRC requested that the Ad Hoc Committee reconcile the positions recommended in the Interim Report with those requested through the budget process. This was done.
Return to
Distributed Computing Support Management Team Home Page.
Page last modified on May 05, 1997 by Philip Baczewski
.