Вы находитесь на странице: 1из 5

ARCHITECTURE FOR KNOWLEDGE MANAGEMENT SYSTEM

Taruna Rawal, Partha Ghosh,Sabyaschi Pattnaik FM University


Balasore, Orrisa
tarunarawal@gmail.com,parthoghoshin@gmail.com

Abstract

This paper intends to project the traits that help to generate, organize and leverage knowledge in social
settings. There is no doubt that in addition to social and psychological aspects, IT if applied in a sensitive
way can be very effective driver for knowledge management. There are various systems that deal with the
representation and dissemination of knowledge; like decision support system, management information
system and expert system and group ware system. This article highlights the substantial similarities
between KMS and conceptual enterprise model. The multilayer architecture is proposed that can be
applied and refined for wide range of companies. The architecture includes an interface to integrate KMS
with existing Information system.

Introduction

Any attempt to define a knowledge management system faces a dilemma. On one hand knowledge-------
seems a self evident term with no need for further explanation. Nevertheless many different definitions
exist .That make it almost impossible to find a definition that is compatible with most existing notions of
knowledge. Knowledge, on the other hand, represents a phenomenon that is very difficult to reflect upon.
While we can speak abut knowledge, any insight into knowledge can be regarded as knowledge itself. As
with language we can differentiate about knowledge and knowledge about knowledge (meta knowledge).
But in the end we can not avoid regressum ad infinitum. For these reasons, it seems to be a frustrating
endeavor to develop a comprehensive definition of knowledge. The pragmatic image of knowledge can
be helpful and is capable of differentiating KMS from traditional Information system (IS).

Proposing knowledge management system would mean the introduction of just another label that
mystifies, and so seems impressive. However in this article the main focus is on aspects of knowledge
that are relevant for its management by machines and that are suited to the deduction of essential features
of system called as KMS.

The architecture

The requirements proposed still allow for a vide variety of systems. As with any other IT artifact, IT
WOULD BE PRESUMPTUOUS TO CLAIM ONE BEST SOLUTION .Here is a outline of KMS which
can act as a suggestion of hoe such a system looks like. It is inspired by the method called MEMO (Multi-
perspective Enterprise Modeling) Frank,1997. Enterprise modeling can be regarded as a medium for
organizational learning and dissemination of relevant knowledge.

The different levels of abstraction that are suggested by MEMO correspond directly to the to the layers of
proposed architecture AS SHOWN IN FIGURE.

Meta Layer Specialized Languages (Meta


models)
Instantiated from

Conceptual Conceptual models


Layer
Generated from

Interface Instances
Layer
The meta-level layer

This layer which we could also called the terminology or language layer, consists of special pear
modeling editors. These editors provide set of modeling languages, like a process modeling language, a
strategy modeling language or an information modeling language. Object models and additional
constraints define the language. Classes of the object model represent the concepts of modeling
languages. The key concepts of modeling business type processes are represented by the classes, process
type, process use, context of process use Input Spec Output Spec and ‘Event’. Process type is an abstract
class that is specialized into two concrete classes: Complex process type and basic process type. An
instance of complex process type may be composed of instances of process use, each of which is assigned
exactly one instances of process type.

Conceptual Level Layer

This layer serves to create, edit and store domain specific knowledge, which is captured in conceptual
models. For this purpose objects are instantiated from the classes that constitute the meta-level layer.
While the conceptual models are rendered with specialized graphical notations their semantics are defined
within the objects that are instantiated from the meta-level classes. These instances still describe types,
not concrete instances. From software engineering point of view this is an important restriction, since it is
not possible to drive concrete instructions simply through instantiations. Usually an instance cannot be
instantiated from another instance. In contrast to me level the content of the conceptual level is usually
created and manipulated by the users of KMS. One would use the various editors to create the objects of
particular classes, and their state by applying the functions provided by the editor. Reuse can be enhanced
with a library of reference models. For knowledge to be an organizational and not only an individual –
asset, it has to be represented in a structure that is well known through the company. It should include
agreements on syntactic rules and semantics both. For instance, it is not satisfactory simply to define a
standard graphic markup language (SGML) document typed for the sales report. In addition the meaning
of the tags should be defined. The knowledge Management systems suggested provides the user with
formal concepts that they can use to describe certain perspectives on the enterprise. It still allows use of
natural language annotations where semiformal descriptions would be feasible.

The interface-level layer

Object –oriented languages distinguish between classes and objects. However these two levels are not
enough. If a concept on the meta-level is specified as a class, it can be instantiated into a particular object
with the name ‘order management’. However, it is not possible to create an instance from an instance, as
would be necessary in order to represent a particular process. Within a traditional IS, no matter whether it
includes an explicit or implicit notion of business process, the user deals with a particular instance that
starts at a specific time and has a specific state. Since the level through instantiation, generating
representations that are appropriate to serve interferences with traditional IS. It is assumed that such an IS
is specified in an object-oriented way. The interface level consists of object models or class definitions
from which particular IS could be instantiated. One also needs functional and dynamic descriptions- such
as message flow diagram, state charts or specific languages to describe work flows. It is not possible to
generate a new version of the interface-level layer whenever the conceptual-level layer has changed.
Therefore, the application of restricted policies for code or model management are recommended. Usually
IS will have been developed before KMS. In this case generating schema information is insufficient.

The Knowledge Management Cycle

For humans the process of transforming data and information into knowledge and then back into value-
added information is a cycle that is natural and on going. Figure 1 depicts this knowledge management
cycle as consisting of four fundamental steps that involves the storage, processing and communication of
information. We begin the discussion of this cycle as it applies to the individual and move on to
discussing the cycle as it applies to small and then large organizations. In each case the methods of
storing, processing and communication information are described and followed by a description of the
progression through the four steps of the knowledge management cycle.

An individual makes his or her way through the world being inundated with data and information from
the environment. To deal with this, an individual uses personal memory as well as notes and paper files
for storing information. The individual’s brain processes the information with possibly the aid of a
calculator or a small computer. Communication of information is primarily internal from a knowledge
management perspective. As individuals we pride ourselves on our ability to learn from our triumphs and
defeats through the effective consolidation of knowledge. As the figure depicts, knowledge consolidated
at the end of one iteration through the knowledge management cycle provides new information that can
be used in yet iteration.

Small organizations of 2 to 20 persons are able to emulate the knowledge management cycle of an
individual with some degree of success. Information and requests received from customers, partners, and
the government is stored within individual memories, in documents and in simple database systems.
Information processing takes place in individual brains as well as at productive meetings where the
strengths and weakness of the individuals are well understood, accepted and utilized. Various small
computer systems and possibly a network server are shared by all. Communication is primarily via ad-
hoc meetings augmented by telephone, fax and email messages when a person is travelling or at home.
Knowledge consolidation by each individual is facilitated by a collective effort to ensure that failure does
not recur for the same reasons and that success can be repeated as often as possible. Consequently, small
organizations are said to be well-oiled, creative and able to move quickly to meet a changing environment
with a high degree of synergy where the value derived from a project can often be greater than the sum of
the individual efforts.

Larger organizations have a difficult time emulating the knowledge management cycle of an individual.
Large companies and institutions receive proposals, queries and other forms of information from a
multitude of customers, channels, partners, government and regulatory bodies. Information is stored in
various formats and locations that include policy documents, filing cabinets, internal process and product
databases as well as external customer and distribution databases, microfiche, audio tape and video tape.
Portions of the information in-flow are processed by individual brains only to be confounded by a
multitude of meetings in which the persons assigned to various roles change from quarter to quarter.
Various computer systems developed over the last ten years process portions of information in silos that
have a difficult time talking to one another for technical and political reasons. Communication is
achieved via a cornucopia of local area network, Internet, mobile devices. Meetings must be scheduled
several weeks in advance for executives and many events must be cancelled and rescheduled due to
conflicts. Knowledge gained at the end of a product cycle is often lost and for this reason failure can
recur and success is not repeated as often as possible. Subsequently, large organizations are said to be
lethargic, lacking creativity and slow to react to meet a changing environment. The chaos that results is
largely due to the ineffective management of organizational knowledge.

Environmental data
Observation
andAnalysis
Problems
Information Opportunities

INFORM ATION
Knowledge Storage Theory
Consolidation Processing Generation
Communication

Approach
Results Methods

Testingand
Application

Figure1
. TheKnowledgeManagement Cyle
Models for structuring Knowledge Management
Knowledge management tools should help the information exchange by three
processes:

1. Knowledge management must help with asking questions


2. Knowledge management must support finding sources of information
3. Knowledge management must help with understanding and the application of
the answers.

Huizing and Bouman (Jentjens (2001)) describe four choices for structuring
knowledge management. The three described processes can be recognized in these
models, although in different ways and intensity.

• In the market model, there is practically no support. Everyone goes his own way
in seeking knowledge and the organisation allows the employee to exchange
knowledge without a structure and with an open gateway to the Internet.
• The organised market model: the architecture of the system has registration
systems, and library systems; there is an agenda system, and e-mail. The
organisation stimulates the knowledge exchange and distribution and there are
intermediaries connecting knowledge seekers and knowledge providers.
• The extended organized market model: the system architecture also includes
interactive systems that can help to make questions explicit.
• The hierarchy: now the system architecture also includes feedback systems that
can give specialized answers (ask the expert) and the organisation has more
extended vocational training programmes, and mentors and coaches.

Conclusion

Based on pragmatic notion of knowledge------ the concept of multi perspective KMS has been proposed.
It is characterized by object oriented architecture as well as generic body of knowledge.

In long run it is desirable to regard KMS as an integral part of a corporate IS . Such integration has
number of advantages. New users as well new employees can move to ‘knowledge level’ to get a deeper
understanding of the corporation. Since KMS offers different perspective on an enterprise, interrelated
through common concepts, it provides a medium for the fostering of discourses between people with
different perspectives. This helps to promote the process of organizational learning, in the sense that
people are supported to understand the position of others with different professional backgrounds. It also
helps in overcoming the common barrier between business people and IT professionals, is regarded as
one of the factors that can block the effective use of computers and communications’.

References

1. T. Davenport and L. Prusak. Working Knowledge: How Organizations Manage What


They Know, Harvard Business School Press, 2000
2. Karl Sveiby. Strategy Formulation in Knowledge-Intensive Industries, International
Review of Strategic Management, Wiley, 1992
3. Karl Sveiby. A Knowledge-based Theory of the Firm to guide Strategy Formulation,
presented at ANZAM Conference, Macquarie University, Sydney, April 12, 2000
4. Yogesh Malhotra. Knowledge Management for E-Business Performance: Advancing
Information Strategy to Internet Time, Executive’s Journal, V.16, No. 4, pp. 5-16,
2000.
5. A. Newell. The Knowledge Level, Artificial Intelligence, No. 18, pp. 87-127, 1982.
6. Michael Schrage, Inteview by Knowledge Inc,
http://www.webcom.com/quantera/schrage.html
7. Maturana and Varela, Autopoeisis and Cognition, London Riedl, 1980
8. Stuart Barnes, 2002, Theory and Practices.