IT Background Of Knowledge Management Systems


Nov 6, 2013 (3 years and 9 months ago)


IT Background Of Knowledge Management Systems
Helena Jilková
Department of Information Technologies
Prague University of Economics
W. Churchill Sqr. 4
130 67 Prague 3
Keywords: knowledge management, information technologies, information portal, document
management, workflow, groupware, text search, taxonomy, New Economy
Abstract: Knowledge Management Systems provide their users with technologies to capture, store,
retrieve, categorize, and provide access to an organization's experience and information.

This paper
describes Knowledge Management Systems from the IT point of view. IT (Information Technologies)
determine significantly the actual ability to share information among the people and other
participants within a defined business process. These technologies, if used properly, can help
organizations improve information utilization and communication among the members of work
he most noteworthy knowledge management technology trend, the rise of enterprise
information portals (EIPs), is also mentioned in the paper.

Portals and the other knowledge
management systems use a variety of techniques to find and display information (document
management, workflow, text search and retrieval, personalization, classifying and mapping
knowledge). These techniques are being briefly described.

What is Knowledge Management
"A little knowledge is a dangerous thing."
Knowledge Management is a modern curse. As a business concept, it covers any coordinated efforts
to maximize a company's performance through creating, capturing, sharing, and applying knowledge
from internal and external sources. This broad concept has been used to describe everything from
filtering information from newspapers, to identifying business opportunities from summarized
business data, to providing initial and continuing employee education. In this paper we will
concentrate ourselves on the technological aspects of Knowledge Management Systems and we will
use a relevant, more narrow definition of Knowledge Management Systems: Knowledge
Management provided by a Knowledge Management System is the use of technologies to
capture, store, retrieve, categorize, and provide access to an organization's experience and
Knowledge has always existed in any organization, has always been spread among the people and has
always been supported by the technologies for information capture, storage, cataloging, access, and
dissemination. These technologies determine significantly the ability to share information among the
people and other participants within a defined business process. These technologies can help
organizations improve information utilization and communication among the members of work
groups. What does it mean? It means that for these processes to become true reality, the proper
technologies have to be applied and people have to manage them properly. The adequate usage of
up-to-date technologies by the people is being called Knowledge Management. The quality of a
company Knowledge Management is one of the competition Critical Success Factors.
"Art and knowledge bring bread and honor."
Perhaps the most noteworthy knowledge management technology trend is the rise of enterprise
information portals (EIPs), also called enterprise knowledge portals, corporate portals, and
collaborative portals. EIPs provide users with a centralized access point to a variety of knowledge
resources via a Web-browser interface. The model for this type of portal comes from successful
Internet applications that are used by both consumers and business-to-business people. The EIP way
of presenting information expects that for the "information users" this type of manipulation with
information is completely natural. Building and usage of EIPs is our common future, nowadays it is
not a matter of our choice any more. (That’s why there is a more detailed description of information
portals concepts and technologies in the part 7 of this paper.)
"Heart is blind but the brain needs to see."
Portals and the other knowledge management systems use a variety of techniques to find and display

document management


text search and retrieval

searching non-text data sources (especially on the Web)

personalization and collaborative filtering, user profiling via agents

classifying knowledge, mapping knowledge

Document Management
"A book that remains shut is but a block."
Database management software organizes and maintains structured data, document management
provides the same capabilities for unstructured and semi-structured data in documents. The
mechanism for managing documents is to create structure in the form of information about the
documents' attributes typically stored in a related "meta-database". Documents can be searched for,
retrieved, and managed based on combination of the "meta-data" and full-content indexing.
Today the term "document" means almost any file that can be stored in paper-based format or
electronically, including text, graphics, spreadsheets, images, program code, and even sound
and video. Many systems capture meta-data about paper documents, where one of the attributes is
that document's physical location. Document management systems are beginning to be
implemented enterprise-wide for general document life-cycle management (creation,
modification, approval, distribution, archiving). However, most systems are implemented within
large organizations in departments where document management provides a strategic competitive
advantage. Industries that were early adopters of document management systems include
manufacturers of engineered products, where the accompanying documentation must be up-to-
date and accurate.
Most commercial document management systems have client-server architecture. The server includes
the document repository, which comprises a document store (often as a file system) in which the
actual document contents are stored. The attributes of documents - the meta-data store - are typically
stored on the server, under a relational database management system. An index engine providing the
full text index of the contents from the document store is normally on the server, as well. The clients
hold the user environment for document access. In a three-tier architecture, the logic implementing
the document life cycle is maintained on a Web or application server, and a separate client, which
might be a Web browser, is used as the user interface.
"Joy which we cannot share with others is only half enjoyed."
Document management systems are also being employed to help users access the mess of available
information on the Internet, organizing it in predefined ways. In fact, document management has
been becoming the integration environment for other technologies such as workflow, e-mail
and groupware. The reason is that for reasonable document system utilization one needs workflow
capabilities with an emphasis on assembly and routing of documents that contain pictures, text files,
spreadsheets, embedded sound, and graphics.
Like many technologies, document management systems have suffered from a shortage of standards
that would make the communication among diverse systems possible. From the user point of view,
the most important set of standards defines markup languages, including Standard Generalized
Markup Language (SGML), HTML, and XML, all of which add structure to documents. SGML is a
powerful language for defining document types. However, SGML syntax makes it complex and
expensive to implement. HTML, a subset of SGML, is easy to implement, but it is not extensible and
has limited capabilities derived from document formatting for display in a Web-browser window.
XML is easier to implement than SGML and is significantly more flexible than HTML. XML allows
information about the data to be defined, whereas HTML is focused on one type data presentation.
"Performing well is the best revenge we can take on our enemies."
There are many document management products available in the market and there are some leading
solutions to the problem but in general, we all are users and producers of documents and within the
context of knowledge management,

we have to manage collaborative features of office productivity software, at least,

we have to be able to select proper tools,

and we have to be able to use them.

"A chain is no stronger than its weakest link."
Workflow technologies manage the movement of information as it flows through business processes.
A workflow application maintain a record of changes in status and the state of the document or
transaction expressing that business process (e. g. purchase ordering). Today, workflow applications
fall into two categories:

collaborative workflow

production workflow
Collaborative workflow addresses project-oriented and team-oriented (collaborative) types of
processes. The focus of collaborative workflow is on allowing workers to communicate and
collaborate within a unified environment. Collaborative workflow can be supported by a specific tool
but its essence belong to project management tools.
Production workflow tools address critical, transaction-oriented processes and are often deployed
only in a single department or to a certain set of users within the department. For example, consider
loan processing in banks. Most of the systems in this category

include a workflow management
engine that controls the flow of information through the processing of each use case or transaction.
This engine interfaces with a database (usually a relational database management system) where the
transaction data are maintained. The leading production workflow solutions also include graphical
workflow mapping for defining the process.
"It is folly to fear what we cannot avoid."
An early step when planning to implement a workflow system is to analyze and define
processes, policies, and practices. Based on realistic definitions of processes, policies, and
practices, organizations must define routes, roles and rules. Routes refer to the order in which
information flows through the steps of the business process. Information can be routed serially (on a
single path from one task to another); in parallel; or conditionally (the routing path is chosen based
on a defined business rule). The routing of information often depends on the business rules and
decisions that are made in the course of the process. For example, in a bank, loans exceeding a
specified limit might need two levels of approval, whereas smaller loans might need only one. The
application of these rules to a specific transaction can be determined by the content of that transaction
or can be based on the explicit authorized selection by the participants in the workflow task. The
information being routed usually is not addressed to specific individuals but rather to the roles
individuals play. For example, in an expense-checking application, the approval request is sent to the
employee's manager, rather than to "Novák". Similarly, group roles are often defined. For example -
rather than sending an application to a specific loan officer, it is routed to the first available officer.
"Everyone is emperor on his own ground."
Two basic architectures of workflow applications are e-mail based (called messaging based, as well)
and database based. Messaging based workflow applications send work assignments through the e-
mail system. Under this approach distribution is easy: anyone with e-mail and workflow application
can receive work, without being dependent on connectivity to a specific server. The lower effect is
the main disadvantage of this approach: first, multiple copies of data flows are created when work is
routed in a parallel fashion. Second, while people are working on an assignment, the system cannot
access the document. The consequence of this is, for example, the fact that a work item that has been
sent to a participant cannot be recalled; the person must finish the assignment and return the work
result before further routing can occur. In a database-based architecture, the work resides centrally
and is much easier to manage. However, all participants must have access to the appropriate server
and remain on-line while participating in the workflow. Performance of such a system is resource
demanding. Web-based workflow combines the benefits of both messaging- and database-based
architectures. The data actually exist in the underlying database on the server. Work assignments are
performed dynamically as URLs (URL = Uniform Resource Identifier). Then users are notified
through e-mail that there is work waiting for them at a specific URL.
"He who would rule, must hear and be deaf, see and be blind."
There are two basic categories of people who need to understand workflow systems: all the users
and all people who advise on business processes (how to analyze and define processes, policies,
and practices and based on definitions of processes, policies, and practices, how to define routes,
roles and business rules).

Text and search retrieval means
"Better half an egg than an empty shell."
Text search and retrieval are always an integral part of an Information Portal or any knowledge
management system. This type of applications has even a long history. Online searching is one of the
oldest business applications dealing with text data - text processing started to be performed on the
mainframes in the mid-1960s. Nowadays, search engines are incorporated into the Portal products.
In addition to the search engines, the effective ranking of retrieved documents is also a crucial
component of text search and retrieval. Most full-text search engines return results that are ranked in
some way. Ranking works according to some measure of how well each document matched the query
(for instance, the number of times that search terms appear in a given document can be evaluated.
Next to it, the location of a term plays a role: matching a search term to a document title, for example,
suggests a better match than matching it to a footnote). Process of ranking is almost a necessary
feature of search engines. Clearly, without some means for distinguishing between exact, close,
and distant matches, the resulting list of page hits would be nearly unusable.
"Like breeds like."
Natural language processing is an area in which big progress has been made during the past few years
in search and retrieval methods. These technologies make use of computational linguistics, which
focuses on modeling the rules and nuances of natural language within the computer possibilities. This
way, search and retrieval queries are being answered on a more and more human-like base.
"A thousand probabilities do not make one truth."
Some content servers and document repositories include capabilities for monitoring information
streams automatically and notifying specific users about new information available in their areas of
interest. The user formulates his/her persistent (standing) query and the system starts to shift relevant
information to him/her automatically. Lotus Notes calls it agents, and others call it a publish-and-
subscribe model. From a user's perspective, it is important to have an easy way of defining new
queries and controlling how often one wants new or updated information. (Because we all have, next
to up-to-date information processing, to perform our normal work tasks.)
"A word to the wise is enough."

Personalization and collaborative filtering features
"Workmen are easier found than masters."
Personalization features let users create a custom work space. It means, for example, that individual
users can create personal information home pages that provide links to the information that most
interest them. (Categories of information can be, for instance, news, sports, and entertainment etc.
Corporate portals, of course, provide the same type of features in a business setting.) In addition,
users can be notified about incoming e-mails, appointments entered in the integrated calendar
application etc. By providing the ability to personalize information portals, companies can ensure that
employees have to deal with only such information that is directly related to their job tasks.
According to the company rules, they can subscribe to internally generated information (such as
reports), and externally generated information, needed to accomplish their tasks.
"You can't win them all."
User profiling is related to authorization: who is allowed what. The authorization has always 3
dimensions – see fig. 1.
Fig. 1 – Three Dimensions of Authorization
"Quod licet Iovi, non licet bovi."
Personalization is often accomplished by using software agents, sometimes called spiders. For
example, agents can be deployed to search through both the Internet and the corporate intranet to find
some data specified through their contents (for example, data on competitors, data on a specific
product etc.).
Collaborative filtering engines represent a very advanced possibility that can also be incorporated
into an Information Portal. These engines make recommendations based on previous users' actions
and preferences. Preferences are being recognized in two ways:

users are asked to rate their choices

the pages users view and choices users make are being tracked
For example, salespeople in one location might be notified automatically of new information that
salespeople in another place found useful - even without their knowing about each others' existence.
It might occur if their interests correlate highly, as observed from their past decisions. Of course,
there are some "pioneer" areas for which collaborative filtering is easier and quite possible: for
example selling books and CDs to consumers.
"There is always a first time."

Mapping knowledge
"Far from the eyes, far from the mind."
A big amount of data is always demanding. For us to be able to convert data into information and
information into knowledge, we have to apply knowledge mapping - topics, taxonomies, and
classifications. Classifying data and documents into taxonomies makes it possible to focus on a set of
related information quickly. It has been proven through usage that a broader, flatter taxonomy might
make more sense than a deep hierarchical one - if taxonomy is too deep, it might be too complex to
be understood and used. It is easier to repeat fishing in a pool than to find the way through a structure
one does not understand. The discussion of text retrieval engines and agents later in this paper,
illustrates the necessity to add some structure to the information pool where the search is to be
performed. Many times, we might need a more explicit summary of the available information.
Sometimes we need to be able to select from pre-existing categories, and sometimes taxonomy
actually makes it easier to locate the relevant documents than full-text searching would do.
"Divide et impera!"
At their most basic, taxonomies are the structures by which users can navigate data repositories and
classification systems. Taxonomies tend to consist of traditional structuring techniques, such as links,
hierarchies and tables. These structures (frameworks) make relationships between different data items
explicit. Typically, both automated and manual processes are employed in the development of the
taxonomy frameworks. Once the framework has been applied, it suggests interactively the relevant
words and phrases that users should use to find answers to questions in predefined areas of interest.
"A flow of words is no proof of wisdom."
Devising a classification system is the first step in building taxonomy. This system groups similar
documents together and puts them into topical categories. Such classification can be generated by
manual or automated means, or a combination of the two.
"We are building a bridge over the sea."
In addition to the categorization methods, standard taxonomy templates for specific industries and
interests are also becoming available, to assist with classifying documents and developing a useful
and usable taxonomy.
"Small choice in rotten apples."
However, the engineering problems presented by more dynamic taxonomies can be significant. New
products and new services appear every moment and the need for up-to-the-moment information
means that new categories need to be created almost instantly. In small-scale knowledge
management systems, which typically depend on a person making decisions, documents can be
classified and new categories can be added manually. In large, complex systems, the necessity of
manual classification and reclassification is not acceptable. If classification is not automated fully,
the knowledge management support must at least make it easy to split, merge, and create new
categories, without influencing documents in formerly existing categories.
"When you are well keep as you are."

Information Portals
"New brooms sweep clean."
The Information Portal (also called the Enterprise Information Portal (EIP), the enterprise
knowledge portal, and the collaborative portal) provides users with a centralized access point to a
variety of information resources, using a Web-browser interface. There is a clear parallel
between Internet portals and the dawn of EIPs. When the Web started to play a prominent role in the
mid-1990s, the number of Web sites was relatively and even absolutely small enough to be managed
in a manual, ad hoc way. However, in the last few years, the number of sites has mushroomed. This
created the need for a "centralized" search for information and for an easily understood and navigable
structure organizing information. Behind Internet search engines there are sophisticated indexing and
classification methodologies powered by automated tools. The need for a similar approach to
accessing information has developed in the corporate world. The reason is clear: the amount of
information to which employees are exposed has grown terribly - "thanks" in large part to the Web.
At the same time there are information technologies available for corporate intranets and automated
information dissemination. EIPs are now being constructed to help employees locate, manage, and
use all this information within the context of their jobs.
Portals also provide us with the opportunity to place structure and order around intranets that might
have become wild and unwieldy. Many intranets have become too complex. No information can be
found easily any more. However, well-designed portals can address this problem.
IT architecture and toolkits for an information portal
"Learn the luxury of doing good."
Using a portal, IT (Information Technologies) can provide a common interface to data from different
applications and enterprise information systems. For this purpose, technologies such as application
integration middleware can be used. In addition, enterprise applications, limited in their
functionality, can be incorporated into a portal. Then, for example, an employee can view SQL
search results or Excel spreadsheets, even if that employee lacks the actual application (program).
Some IT suppliers offer toolkits to help create these read-only programs.
Portals use the same variety of techniques and means to find and display information that has been
mentioned in parts 2 – 6 (text search and retrieval means etc.).
Fig. 2 – IT Architecture of an Information Portal
Information sources for an information portal
"Knowledge is proud that we know so much."
Information sources for an information portal are following:

structured data: on-line transaction database produced by ERP applications, Data

unstructured data: Web documents, e-mail, correspondence, product specifications, product
plans, legal contracts, all other Word documents, advertisements, purchase orders, paper
invoices, research reports, audio files, video files, graphics and graphic design files etc.
All these sources are data sources that can be converted into information only through adding some
structure to unstructured data and through applying content-based search for both categories of
the data sources.
"Every man as his business lies."
There is another most current “curse” and “marketing label”, next to the Knowledge Management -
Business Intelligence. The most significant difference between these two concepts (besides the fact
that, if they are sold separately, they are paid twice) is probably the aspect that Knowledge
Management is primarily focused on unstructured data and Business Intelligence on the
information derived from structured data (Data Warehouses, Executive Information Systems
etc.). In an information portal, both Knowledge Management and Business Intelligence, come
again together.
However, we have to warn you: no IT can create information, technologies only help to add some
structure to both the data and the process of information usage. Garbage in, garbage out!
Who will create, maintain and manage your information portal
"Mind no business but your own."
There are many companies for which, because of their specialization or size, creating and running
their own information portal may seem quite a burden. At the same time, all companies need to profit
from such a portal because otherwise their possibilities to compete would decrease. For such
companies that prefer to have their portals maintained and managed by a third party, outsourcing
options are becoming available. Rather than install and implement information portals in-house,
some companies are using Application Service Providers (ASPs). ASPs host applications and charge
a fee to provide access and use of the applications. The fees are typically based on usage. This
approach allows smaller, or specialized companies to use advanced applications but without incurring
the associated implementation and maintenance costs. Application Service Providers make
applications available on a subscription basis. Instead of being sold (or licensed) as a normal software
product and installed on a computer system of that customer, the application software is provided as a
service that is hosted at the ASP's data center. The application then can be accessed remotely by the
In addition, for a specific target group of companies, a vertical portal can be prepared. (A vertical
portal is one that is tailored to the needs of a particular industry, such as banking or energy, or to the
needs of a specific type of companies - for instance, consulting companies.)
"Never cross a bridge until you come to it."

“Where, why and what” of a Knowledge Management System
Where is knowledge in a company?
"In all labor there is profit."
Where is knowledge in a company? In the heads of people and in structured and unstructured data
maintained by various technologies. When we want to manage knowledge, there are two core aspects
related to this "where":

managing existence and location of knowledge

managing privacy and security of knowledge
Existence and location of knowledge are captured in meta-data. Meta-data is essentially data about
data and typically describes which process or program originated the data, how it is structured,
where it is located, how it is to be accessed and who is authorized to access it. Meta-data have to
be and are supported by the knowledge management means.
"Keep yourself from opportunity and God will keep you from sin."
In general, for e-Business security is critical but discussion of security topics such as firewalls,
gateways, and bastion networks has not been attempted here, because of limited space. Security
technology is called upon to perform several functions in e-Business transactions, such as
authentication, confidentiality, secure delivery, and privacy. Within the knowledge management
context, authentication and confidentiality are the most important security requirements.
Authentication is the process establishing an association between a transaction or message and
a person - or, more properly, a "principal" - because authentication increasingly will involve agents
or other software realized, institutional representatives, next to human users. In general, it is not the
particular user but an institutional capability that is implied in an authentication. Not Mr. Novák but
Mr. Novák's actual role in the company is being authorized.
Authentication methods are based on something users know (such as a password), something
users have (such as security tokens or smart cards), or something users are (biometrics).
Confidentiality ensures the privacy of a transaction or message. Confidentiality on computer
systems is normally obtained by restricting access to the data. This restricted access requires some
form of authentication for access. Confidentiality over public networks is obtained by encrypting the
transmitted data. An encryption algorithm transforms plain text into a coded equivalent, known as the
cipher text. The coded text subsequently is decoded (decrypted) at the receiving end. The encryption
algorithm uses a key, which is a binary number. Two types of algorithms are in use today: shared
single-key (also called secret-key, symmetric key) and public key (or asymmetric key). In the shared
single-key method, the same key is used to encrypt and decrypt the message. Public-key encryption is
based on two keys: one to encrypt the message and another to decrypt it. The algorithm is not
symmetric, so knowing the public encryption key does not help in being able to decrypt a message.
Using these methods, knowledge management can be implemented, for example, on a remote
server and from the organizational point of view even outsourced (that means that technological
matters are being handled by a third party organization). Hardware and software means necessary for
making this process completely safe are currently available (for example, the most popular way in use
today to protect sensitive information such as payment data uses the Secure Sockets Layer - SSL -
protocol which has become a de facto standard).
"It is easier to hurt than to heal."
Why do you need Knowledge Management?
"Necessity is the last and the most powerful weapon."
The quantity of data that has to be absorbed, structured and used in any organization, has been
growing terribly. At the same time, we cannot escape from this situation, we have to manage it. That's
why we need to be "upgraded" for this task continuously and we have to use the proper tools
properly. If recent and current situation is any indication, the need of knowledge management will
explode in the next few years, largely due to the increase prominence of EIPs (Enterprise
Information Portals). The Internet and the Web will continue to have a profound effect on the
overall market-place. The adoption of EIPs will accelerate the penetration of knowledge management
systems within organizations. Because of the common familiarity with consumer portals, a greater
number of users will now expect their corporate intranet to offer similar capabilities (such as search
engines and automatic document summarization). Along the way, this trend

should help
organizations transform unwieldy corporate intranets into true knowledge systems. Electronic
document management products are becoming

more useful and necessary in business because they
are more and more business-process centric. Web-related technologies will accelerate the
deployment of document management solutions throughout organizations. Use of document
management systems will be no longer restricted to writers, editors, and document specialists.
Instead, most office-working employees will have to be using the capabilities of their networked
document repositories, to manage, control, and access to information as part of their daily work
activities. Workflow will continue to grow in usage and the embedding of workflow capabilities
into corporate applications will continue. The role of business consulting related to knowledge
management has been growing: the quality of knowledge is dependent on the quality of analysis
and definition of business processes. In particular, workflow products will be sold primarily
through systems integrators as a part of a business solution.
"Not possession but use is the only riches."
Extranets have started to play a much larger role in knowledge management. Knowledge bases
built on information drawn from partners, customers, and suppliers, and/or found on the Internet,
have become essential for companies to compete. Having stressed the role of extranet information,
we can conclude that even a small size of company does not block the necessity to jump into the
knowledge management train. Nowadays, we all need to be continuously upgraded for the
current e-Business level. The integration of business processes across organizational boundaries
adds new urgency to this need.
What does it mean, to apply Knowledge Management?
"Attack is the best form of defending."
The "knowledge management" activities can be divided into two types:

educational activities

activities related to the knowledge management system, planned or existing in a company
Educational activities cover

ad hoc training the people (for example, for using a specific software tool - Lotus Notes for
document management and groupware)

continuous "upgrading" the people for the current level of knowledge management systems
(for example in the form of Internet and intranet based training, tailor made courses,
workshops and seminars)
The main activities related to the knowledge management system, planned or existing in a
company, are following:

selection of knowledge management tools

creating hardware and software infrastructure necessary for a knowledge management system

projects for the implementation of a knowledge management system and its integration with
the other parts of a company's information system

establishing rules and procedures for the contents of the "knowledge database", creating
meta-data reflecting contents, location and other attributes of information (information has to
be selected, order on information resources has to be imposed, the structure to ill-structured
information has to be added, proactively capturing information that might be useful in the
future, is to be anticipated)

"knowledge database" is to be filled once using existing data sources (data has to be prepared
and converted using rules and procedures mentioned in the previous paragraph)

establishing authorization rules for using the knowledge management system

maintaining "knowledge database"

using the knowledge management system

evaluating the usage, costs and benefits of the Knowledge Management System

improving the Knowledge Management System, based on the evaluation
"Everything has its price."
/Bloor/ Bloor, R.: eRoad, Challenge House, United Kingdom, 1998, ISBN 1 874160 39 2
/Minoli/ Minoli, D., Minoli, E.: Web commerce technology handbook, New York, MacGraw Hill,
/Peppers/ Peppers, D., Rogers, M.: Enterprise One To One, DoubleDay, NewYork, 1997, ISBN 0-
/Phillips/ Phillips, D.: Principles That Work At Work, IEEE Computer Society, Sheridan Books,
California, USA, 2000, ISBN 0-8186-8300-7.
/PWC/ Technology Forecast:2000, PricewaterhouseCoopers Technology Centre, California, USA,
2000, ISBN 1-891865-03-X.
/Online sources/
/Online sources/
/Online sources/
/Online sources/