UN
implementing an ambitious integrated management
information system
Mandated by the General Assembly's resolution
43/217 of December 1988, the United Nations
has been developing an integrated management
information system (IMIS) for a client-server
environment. When fully implemented, the IMIS
will have replaced numerous independent systems
at eight duty stations, including the ESCAP
secretariat in Bangkok. The level of effort
and the human resources needed to design, develop,
and implement the system were initially heavily
underestimated, but today the Organization is
starting to see the first results of its US$
75 million investment. IMIS has already streamlined
many administrative procedures for the benefit
of the organization.
Perhaps the widest scope
The United Nations Integrated Management Information
System (IMIS) is in its category one of the
most ambitious systems ever implemented by any
organization. From the beginning, it was clear
that no off-the-shelf product was capable of
meeting the requirements of processing of, and
reporting on, administrative actions at all
major duty stations.
IMIS integrates three
functional areas:
Human Resources - Includes applicant management,
roster and skills management, personnel actions,
post management, entitlements, insurance, payroll,
and leave and attendance
Accounting and Finance - Includes funds distribution
and execution, accounts payable, accounts receivable,
general ledger, disbursements, investment accounting,
and banking and cash management
Support Services - Includes purchasing, suppliers'
roster, cataloguing, travel and travel claim
processing, stock management, and property management.
IMIS incorporates a new, comprehensive, flexible
account code structure that captures data required
to perform general ledger, treasury, and other
financial management functions. Additionally,
it expedites accounting for other activities
that have financial implications such as salary
advances and travel claims.
Thorough integration
IMIS provides full integration among all the
above applications areas. Full integration means
that information on persons, account codes,
allotments' balances, vendors, reference tables,
organizational structure, etc. is entered only
once and is used by all system functional applications.
For example, a staff member's promotion or processing
a procurement automatically updates related
accounts. This integrated database approach
eliminates redundant data entry, inconsistency
of data, and the need to reconcile data.
In IMIS, data are entered once, but can be
used by different offices for different purposes.
For instance, a personnel action such as updating
dependants affects other benefits. If there
is no information on dependants,, there can
be no education grant and no home leave travel
for dependants. Likewise, when the payroll portion
of IMIS is implemented, the staff member cannot
be paid without a valid post and a valid appointment.
Because of this integration, IMIS actions must
be made in a timely manner. To facilitate this,
the system prompts users, edits data, and generally
makes it easier to enter all the relevant data.
IMIS validates transactions as they are processed.
As a result of this validation, many errors
are avoided or detected immediately and corrections
can be made before the transactions are processed.
IMIS prompts the user for a correction whenever
possible. IMIS consistently applies entitlement
rules in all necessary modules of human resource
management and payroll. Also as financial transactions,
such as obligations, are processed, IMIS immediately
checks for the availability of funds. This feature
prevents overspending and enables the responsible
offices to identify funding problems in a timely
manner. IMIS also has an on-line help feature;
for example, there is an on-line list of valid
account codes to choose from.
IMIS provides on-line query reporting facilities.
Because transactions are validated and fund
availability is assured, the IMIS reports accurately
reflect the state of the duty station as of
the time of the report.
Security features are
important
IMIS contains inter alia: personal
data, a comprehensive view of the UN financial
position, and sensitive procurement data. Thus,
restricting access to IMIS is important. The
security philosophy of IMIS is to limit access
to data to those persons who, because of their
function, need such access. IMIS limits the
ability to make changes to the data base to
those authorized to do so. IMIS makes it possible
to monitor who performed what actions. Access
to IMIS is controlled through the use of identifiers
and passwords. IMIS maintains tables of passwords,
authorized users, and audit trails. IMIS records
and reports on attempts to breach security.
IMIS audit trails will be used by management
for analysis of the workload and procedures,
and by internal and external auditors.
Impact of the initial
implementation
After the first two years of operation of
the first part of the system at Headquarters,
and with the completion of the development of
Release 3, related mainly to financial applications,
the impact of IMIS has proved already far greater
than initially expected. In the end, IMIS will
lead to the greatest changes in the operation
of the Organization since its inception. Procedures
are being modified, accurate data on staff and
post incumbencies are now immediately available,
and financial data will soon be available. Most
important, entire working methods are being
changed. Greater decentralization has been achieved,
and monitoring tools for these activities are
being made available to the Department of Administration
and Management. A more global approach to administrative
procedures involving all parties concerned has
been established because IMIS, by its integrated
nature, has shown the interaction and dependencies
of all activities involving these user groups.
From a technical point of view IMIS continues
to operate at Headquarters with practically
no interruption and the technical architecture
for both the system and the recently introduced
reporting facility are in line with industry
standards.
IMIS Releases 3 and 4
IMIS Release 3 (Financial transactions), which
was introduced in May 1996 at the Headquarters,
is by far the largest and most complex of the
entire system. Its implementation has introduced
massive changes in the way financial transactions
are performed, and major difficulties are likely
to arise. In particular, the change in the way
that transactions have been manually processed
for more than forty years will be challenging.
The Organization has reviewed in detail hundreds
of processes to determine the most appropriate
way to reengineer and standardize procedures
and rationalize the work flows. In some cases,
the need for consistency and the requirements
for additional data not recorded so far has
led to procedures that may appear cumbersome
to users. That problem will be alleviated by
user training and manuals that integrate IMIS
operational procedures, including manual procedures
outside of the system, with financial and support
services operations.
The most important benefits of Release 3 will
derive from the timeliness in the availability
of financial information, both on line and as
printed reports available to senior managers,
and from the funds sufficiency checking and
control that will provide certifying officers
with the ability to ensure that commitments
are not made without appropriate funding.
Release 4 comprises payroll, time and attendance
reporting, rental subsidy and personal insurance.
The development of these applications will rely
primarily on the information contained in the
database and supported in the first three releases
of the system. In parallel, a series of policy
issues, together with an analysis of local procedures
either internal to each duty station or dependent
upon local factors, such as the exchange of
information with local banks, is being reviewed
in order to better define the scope of the development
work.
Lessons learned so far
The implementation of the first release of
IMIS in September 1993, comprising personnel
applications, confirmed the need for the system,
the benefits that the Organization and Member
States will derive from it, and the validity
of the technological choices that had been made.
It also revealed, however, that the Organization's
strategy for the completion of the software
construction phase of the system and for its
successful implementation worldwide had been
inadequate. In early May 1994, a thorough reprogramming
and rebudgeting of the project were therefore
undertaken with the aim of redefining the level
of resources that will be required for its completion
and of determining the resources that can be
internally reassigned to the project, all for
the purpose of providing Member States with
all the necessary accurate information on the
expected future costs of the project.
The operations of IMIS applications already
implemented have also confirmed the need to
ensure that the system, once completed, is properly
maintained. Maintenance will ensure that the
initial investment is not wasted, the value
of information technology having been confirmed
as one of the most important requirements for
modern management.
The implementation of the new releases at
Headquarters, and the implementation of the
system as a whole at offices away from Headquarters
and peace-keeping missions, present many challenges.
As is the norm in this type of undertaking,
and taking into account in particular the fact
that the introduction of Release 3 will bring
far more extensive changes in United Nations
administrative operations than the first two
releases, some initial confusion and resistance
on the part of users can be expected. It is
expected that some programmes in the system
will prove inadequate and will need to be modified
very rapidly; more difficult changes will require
the development of workarounds pending their
coding into the system and procedures will need
to be adjusted.
Conclusions
The impact of the IMIS project and of the
introduction of the system goes far beyond what
was originally expected because new administrative
procedures, new management behaviour, new technologies
and new information technology strategies had
to be developed and introduced. It has been
a difficult process and major difficulties will
still have to be faced in 1997. It can, however,
be stated that the Organization as a whole will
benefit from this experience and that, also
because of this major undertaking, the Organization
will soon be able to be more responsive to Member
States requests for increased efficiency, accountability
and responsibility. The project, however, should
be seen only as the first step towards a more
continuous and steady effort to improve efficiency
in the United Nations through advanced technological
innovations.
Developing any large client/server system
has its pitfalls, but the frequently changing
nature of United Nations administrative procedures
itself was an impediment to the IMIS team. To
make future changes of the system easier, the
design had to avoid hardcoding of the administrative
rules and parameters in the programs. Another
difficulty was that the old mainframes were
to be maintained online, at least until IMIS
is up and running in all eight locations. Some
functions, such as payroll, will be performed
in parallel - on both the old system and by
IMIS - until IMIS is proved reliable.
The Integrated Management Information System
(IMIS) was a winner of Datamation's 1994 Client/Server
Solutions Award. IMIS has been adopted for the
use by the headquarters of the International Labour
Organization (ILO), the United Nations Children's
Fund (UNICEF) and the United Nations Development
Programme (UNDP).
If you would like further information
about IMIS, please contact Mr Zoltan Nagy, IMIS
Coordinator at ESCAP, tel (66-2) 2881251, e-mail
nagy.unescap@un.org.
Technology
featured: client-server computing
The IBM Dictionary of Computing defines
client-server computing as follows: "In TCP/IP,
the model of interaction in distributed data
processing in which a program at one site sends
a request to a program at another site and awaits
a response. The requesting program is called
a client; the answering program is called a
server".
The simplest form of a client/server application
has only two tiers: a client program and a server
program which are connected via a network. In
a typical application, the client program manages
the user-interface portion of the application,
validates data entered by the user and dispatches
requests to server programs. Server programs
most commonly receive requests from client programs,
execute database retrieval and updates, manage
data integrity and dispatch responses to client
requests. Functions that require high processing
power are usually programmed into the server
process. The server commonly manages also shared
resources such as databases, printers and communication
links. The client process is often called the
front-end of the application, whereas the server
process forms the back-end.
From monolithic to three-tier
models
PC and mainframe applications are representatives
of monolithic architecture, where all
tiers of an application are woven together and
usually work as one piece.
Most client/server applications are two-tiered,
where application logic is split between two
separate processors, one a client and one a
server. The presentation is usually at the client-end,
but it could be partially also in the server;
data are usually housed in the server, whereas
the functionality could be in either end or
be split between the client and the server.
In three-tier architecture, the application
is separated clearly into three distinctive
modules, namely the presentation, the functionality
and the data.
Distributed computing
Distributed computing is basically client-server
computing on a wider scale. Data are not located
in one server, but many servers. These servers
might be at geographically dispersed areas,
connected by wide area network (WAN) links.
(LAN Times - Encyclopedia of Networking).
Such systems are often called enterprise
networks because they join the formerly autonomous
computer systems in separate departments of
an organization. They may consist of multivendor
computers, operating systems and applications,
all made available transparently in a common
network.
*****
For technological details on client-server
and distributed computing, please refer to numerous
books and articles available in bookstores and
on the Internet. One of the better studies the
Editor found on the Internet (http://www.finance.gov.au/pubs/client_server/toc.html)
was made in January 1995 by the Information
Exchange Steering Committee in Australia: Client/Server
Computing in the Australian Public Service:
The Next Wave? The copyrighted publication
contains a balance of technology, management
issues and case studies. It deals among other
things with organizational requirements of application
development and management, pros and cons of
the technology and the future of client-server
computing. "People issues" and "business cases"
are particularly well covered.
To
centralize or distribute?
The client-server data processing model allows
a number of users to access common server resources,
which are centralized. Distributed computing
on the other hand means that data, like other
resources, are controlled in various locations
of the organization. Organizations may thus
find themselves at the same time creating centralized
server databases and providing access to decentralized
data holdings.
Distributed systems became possible with the
movement towards open systems and the development
of network technology, which provides fast,
reliable and cost-effective communication between
computers.
Distributed systems offer benefits of client-server
computing in accessing data by several simultaneous
users, while local managers can control data
that they are familiar with. Distributed systems
are thought to provide a better protection from
loss or downed systems than those depending
on a single server.
Distributed systems however are not all that
simple to manage. Keeping data synchronized
is especially difficult. Distributed systems
often only integrate existing systems through
fast communication links, which can create chaos
compared to newly designed systems.
An ordinary user manipulating the system through
client software probably does not care which
technologies he is using indirectly, where the
processing takes place or where the data are
located. But the boss and the IT manager will
definitely learn about it if the system hangs
up, crashes, performs too slowly or loses and
corrupts data when the user expects to complete
mission-critical functions.
What
to do with existing applications
It is not always easy to get the development
of fresh large-scale information systems rolling.
Many steps are involved, including an analysis
of the business environment and the need to
improve information systems, calculation of
expected benefits, selection of development
strategies (e.g. turn-key vs in-house), selection
of hardware and software environments and development
tools, scoping of the application, possible
modularization, master design, prototyping,
development itself, testing, and implementation.
In spite of systematic development, most systems
seldom become complete. After the initial development
cycle is over, bugs are invariably found and
fixed. No doubt alert users initiate improvements,
which, if fundamental, will be implemented by
using the same development methodologies as
were used for the system proper.
Even the best systems will finally come to
an end of their life cycles, usually when they
become technologically outdated or their capacity
proves insufficient. At that point the organization
needs to develop a better system and move its
business to the new application. Whether the
transfer is from a mainframe or from a PC environment
to client-server, the migration is a difficult
task that requires careful planning so that
business activities can be performed uninterrupted.
A major commercial provider of servers and
host computers, Sun Microsystems Inc.,
emphasizes that it is imperative that there
first be an overall vision that links the [organization's]
future information technology to its business
drivers and processes (see http://www.sun.com/
sunservice/MOVING/moving_intro.html). An important
success factors for the migration is an executive
buy-in and some consensus in the organization
about the direction of information technology.
Migration becomes easier if it is broken down
into manageable pieces:
- Identifying key business
processes of the enterprise
- Selecting an application
architecture
- Identifying and scoring
applications with major impact to those business
processes
- Choosing migration options
with attention to factors such as skills and
tools which enable the process
Sun divides migration options into
five major "redo's":
Refront - Addition of a new user interface
to an existing legacy application.
Replace - Swapping of a legacy application
for a packaged application on a new client-server
platform.
Rehost - Porting of a legacy application
onto a new platform with no change in functionality.
Rebuild - Migration of an entire application
to full 3-tiered computing.
Redesign - Re-engineering the surrounding
business process along with rebuilding the legacy
application into a 3-tiered peer computing architecture.
Development of a client-server application
should not be an aim by itself. An analysis
of the organization's business process might
reveal better ways to improve productivity of
the information system than a large-scale migration
to client-server. However, it is not unusual
to use the implementation of client-server architecture
to force changes in the business processes.
From the previous article it is evident that
the United Nations chose to redesign practically
everything when it developed an integrated management
information system. Without making any judgement
on whether the Organization's priority was on
the rethinking of administrative and financial
procedures, or on the development of a modern
management information system, the final outcome
appears to be a happy one. The Organization
is also in the process of improving other parts
of its information systems, most notably the
economic and social information system, which
will involve among other things refronting of
some of the existing databases.
More
popular than you may think
Client-server computing is already enormously
popular. If you ever used Internet, you have
used client-server applications. For instance
when you are browsing the World Wide Web, you
have to operate client software, such as Netscape,
Internet Explorer (graphics-based) or Lynx (text-based),
which are talking to a Web server you are connected
to. Other client software, such as newsreaders
and e-mail software, are part of the standard
Internet arsenal as well. When using them through
your client, you need to know basically only
one thing about Internet servers, their addresses,
e-mail addresses and URLs, which you have to
type in your client software.
The next issue of the Newsletter will
dwell on Internet issues. The future of the
Internet seems bright, although for the time
being increasingly congested.
|
