2. Knowledge Management
2.1. Introduction
Shannon [
1
] and Weaver [
2
] introduced the concepts of data and information. Timpson [
3
]
resume their ideas; he considered the data as a representation of a fact isolated, out of context;
and information as collection of data structured in one context (data with meaning). The notions
of data and information were incorporated into the DIKW Hierarchy, presented by Ackof.
The DIKW Hierarchy (see Figure 1) presents a hierarchical chain in that each concept adds
value to the previous; in the lowest level is found the data; one higher level, information; then,
knowledge and in the superior level, the wisdom [
4, 5
].
Cornella [
6
] says that transform information into insights produce knowledge in the people
minds, generating structure and making sense [
7
]. The knowledge is a “appropriate and
persistent response to a given input” [
8
] and his value is associated with his use. Finally, “if the
knowledge is subjected to judgment value and it gives an ethical context, becomes wisdom " [
9
].
Díaz et al., [
10
] define the knowledge as: "the mixture of cognitive and contextualized beliefs,
perspectives, judgments, methodologies, information, experiences and expectations about an
object, which is adapted and empowered by the people mind". Martínez Rueda [
11
] proposes
that the knowledge provides the basis for generate a critical view of the environment. That is,
the environment can be understood and interpreted through to knowledge.
Knowledge Management (KM) is a process to through of the which the organizations can
discover, use and maintain the knowledge, for align it to the business strategies and to obtain
competitive advantages [
11
]. Perez et al., [
12
] define the KM as an emerging discipline for
exploit the knowledge generated to achieve the organizational objectives and optimize the
process of decision making".
2.2. Knowledge management views
For Gómez-Vargas et al., [
13
], the Knowledge Management is supported by four components:
people, processes, content and I&C (information and communication) technologies [
14, 15,
16
]. These components match with the considered in [
17
]: people (roles and responsibilities,
motivation, incentives), technology, processes and content. The notion of "content" is similar to
“content representation” and its methods [
18
]. Milton [
19
] define the follow KM drivers: roles
and responsibilities; process; governance (policies, metrics and incentives) and technologies.
According to Zytniewski [
20
] and Meier [
21
], in order to be successful in managing knowledge
in organizations, it’s necessary to consider the organizational, social and technical aspects.
The organizational aspects are also present in Basit et al., [
22
], Cumberland and Githens [
23
]
and Wahda [
24
], who consider it is relevant to include organizational culture, also called
“institutional aspects” by Sobolewska et al., [
25
].
Basit et al., [
22
] presented an investigation, based on Brooke et al., [
26
] and Park et al., [
27
],
where it is concluded that employee trust is a prerequisite for sharing knowledge. Saavedra et
al., [
28
] consider culture as a relevant factor.
For Wigg [
29
], an organization capable of creating, acquire and transfer knowledge, must
hire people willing to learn. Nonaka et al., [
30
] also consider some aspects that the organization
must support in the knowledge conversion process: organizational commitment, autonomy at
all levels, flexibility, redundancy (repeated knowledge throughout the organization) and demand
for diversity of personnel and knowledge.
In this paper, the components or dimensions are referred to as "views". Straccia et al., [
31
]
resume the possible views, based on Gomez et al., [
13
], Barceló [
32
], Meneses [
18
] and Milton
[
19
]. The fundamental views for KM are people role and responsibilities, organizational culture
(and others organizational topics), process (or activities), governance, knowledge representation
and technologies.
3. Knowledge management architecture
3.1. Introduction
The architecture constitutes "a relatively small and intellectually accessible model of how the
system is structured and the way its components work together" [
33
] and show early decisions,
which will have a profound efect on all subsequent software engineering work and on the
success of the system as an operational entity [
34, 35
]. An architecture is a system model
within a specific context, representing the components necessary to develop the system from
a particular perspective or point of view. A system architecture includes the structure of the
core components, their characteristics or properties, as well as the relationships and interfaces
between the components and external entities [
36
].
3.2. KM architecture
According to Medina et al., [
37
], the architectures for Knowledge Management Systems can be
divided from two approaches: classical architectures and proprietary architectures. Proprietary
architectures are structured with a strong component of agents and are related with proprietary
software with their consequent integration and dependency. Classical architectures are
considered generic and can be applied to all types of Knowledge Management System and their
architectural style is N-Layered, organized by layers that communicate to complete the required
functionality; each layer has a well-defined scope and functionality.
Woods et al., [
38
] present a seven-layer KM architecture with these layers: application,
interface, KM services, taxonomy and knowledge maps, process and information management,
infrastructure and data source. This model is used for Lawton [
39
] too and is called Ovum.
Tiwana [
40
] proposes a new seven-layer KM architecture: interface, access and authentication,
collaborative filtering and intelligence, application, transport, legacy integration middleware
integration and repository.
One of the main knowledge management models is the Kerschberg Model, which proposes
a technological architecture for knowledge management defined by three layers: knowledge
creation and presentation, knowledge management and data source [
41
]. Zavala Zavala [
42
]
and Jofré et al., [
43
] present other models of technology integration architecture based on the
Kerschberg model, with the same layers.
Finally, Moscoso [
44
] proposes a hybrid infrastructure for intelligence, business analysis and
knowledge management based on an educational data warehouse (EDW) and an enterprise
architecture repository (EA) that allows the digitization of knowledge for the visualization and
analysis of organizational components. The main components of its infrastructure proposal are a
software tool for modeling organizational components, an EDW (Educational Data Warehouse),
OLAP tools, EDM software and visualization tools. However, unlike the other architectures, it
does not use a layered model, which is the scope of this work.
3.3. N-layered architecture
In the previous section diferent architectures were presented, identifying three layered models:
Ovum, Tiwana and Kerschberg. In this section these models are analyzed and compared. When
knowledge management architecture models are analyzed comparatively with 3-layer software
programming models, some similarity can be found. The 3-layer software architecture models
have the following layers: User Interface (UI), Application Logic and Database (or Data) [
45
], as
shown in (see Figure 2).
The 3 models match in proposing a presentation layer; although in Ovum, it is subdivided
according to the characteristics of the applications, reserving a first layer for applications in
general, while proposing a second layer exclusively for the knowledge portal.
Regarding the lower-level layers, all models also coincide in the existence of a repository
and data source layer, although Tiwana incorporates some technical aspects of networks and
hardware as a transport layer and Ovum proposes an infrastructure layer, aspects that are not
considered in this paper because they are considered out of its scope [
46
].
For this work, Kerschberg’s nomenclature is used: Presentation and Repository, respectively.
In the models presented, there are some diferences with respect to the intermediate layers
(see Figure 3). Ovum proposes the following layers: Knowledge Management, Taxonomy
and Knowledge Maps and Process and Information Management. Tiwana opts for two layers:
Collaborative Filtering and Intelligence and Application. Finally, Kerschberg incorporates
everything in a single layer.
Considering that Ovum’s architecture is the most detailed, this paper considers its 3
levels for the present work. In addition, Tiwana innovates with the proposal of an access and
authentication layer that is later used in [
47
]. The resulting model can be seen in (see Figure 4).
4. Knowledge management technologies
4.1. Introduction
Without attempting to constitute a systematic review of the literature, but following an orderly
search process, a search was conducted on various knowledge management articles that propose
technologies. A search is performed on the Association for Computing Machinery (ACM), the
Institute of Electrical and Electronic Engineers (IEEE), Directory of Open Access Journals (DOAJ)
and Science Direct. For the first sources the search is performed on the year 2019, while in
Science Direct the search corresponds to the years 2017 and 2018. The findings of these searches
are presented below.
4.2. Technologies
For Jin-Feng et al., [
48
], the information technology should include "Internet, intranet, extranet,
mobile network, browser, data storage, data mining technology, software agents, network
community, cloud platform" and open platforms such as Baidu, Wikipedia or Zhihu, which are
referred to as wiki in this work.
Paschek et al., [
49
] present a model of KM parallel to the Business Process Man-agement
(BPM), highlighting the importance of managing knowledge in parallel to business management
to adapt to the constant changes in the organizations. The authors propose the following
technologies: Big Data, cloud apps, Datawarehouse and Business Applications.
Córdova et al., [
50
] show the implementation of KMS and states that "technology adapts
to organizational structures and needs, providing support tools for knowledge management,
through the interaction and use of collaborative tools". The proposed knowledge management
support tools are classified in Table 1.
Wamuyu et al., [
51
] say that "to support knowledge sharing among the County Government
employees, counties should deploy knowledge sharing systems such as collaborative systems
and knowledge repositories to allow the employees to acquire tacit and explicit knowledge from
each other". However, it does not provide significant technology-related details.
Roblek et al., [
52
] show an analysis about the smart city knowledge management. They
present diferent KM views, including the technologies. The work recommended the data
collection and exchange through information technologies components and devices which
are connected on the internet or wireless, big data management and artificial intelligence
applications. Then, the authors propose use platforms such as e-health, e-gov, e-learning,
e-social connections for seniors, e-mobility.
Bongku et al., [
53
] compare Knowledge Management Applications: OpenKM, Alfresco and
XWiki, based in several criterial. Alenezi et al., [
54
] discuses about the use of ontologies for KM
specific to soft-ware security. Although it shows ontologies, most of the paper compare specific
technologies like OntoCheck, RDF Triple, OOPS.
Liu [
55
] researches on Knowledge Management Technology of Aerospace Engineering Based
on Big Data and shows the Hadoop architecture. Hadoop is a distributed system infrastructure
developed by Apache Foundation and implements a Distributed File System. The author shows
a framework of Knowledge Management with activities and technologies.
Hibbi et al., [
56
] show an expert model with a Bayesian Network method. This model proposes
process knowledge to combine three categories of knowledge: Declarative, Procedural and
Conditional and shows an Intelligent Tutoring System (ITS) that includes Instructional Module,
Expert Module, Implicit and Explicit tutors, Pedagogical Agents and Collect Data resources.
Tseng et al., [
57
] show a study that provides valuable findings regarding the possible causal
relationships among social media use, Knowledge Management Capability (KMC) and supply
chain agility. The results indicate that social media use have an indirect influence on supply
chain agility through the mediation of KMC.
Other papers use some aspects of KM and propose some tools: Ismail et al., [
58
] analyze the
social presence in digital community and propose use of Agent-Mediated Personal Knowledge
Management and Ferchichi et al., [
59
] propose an ontological rule-based approach for software
product lines evolution and present use of Protégé OWL and SWRL.
Cerchione et al., [
60
] provide a literature review of how knowledge management is applied
in SMEs. It analyzes diferent articles and present the followings: tools that support the creation
phase, tools that support the storage phase, tools that support the transfer phase, practices that
support the creation phase, practices that support the storage phase, practices that support
the transfer phase. The result is a detailed taxonomy presented in Table 2 where the tools and
practices that support each of the phases of the KM process are included.
A few authors present technologies with some architectonic relationship. Xiaohua [
61
]
recover the seven-layer architecture KM and propose some technologies for three layers:
• in the collaborative filtering and intelligence layers propose intelligent agent tools, content
personalization, querying, indexing and metatags;
• in the application layer the works proposes skills directory, yellow pages, collaboration
tools, video conferencing, digital whiteboards, electronic forums, DSS (Decision Support
Systems) and GDSS (Group Decision Support Systems); and
• in the repository layer proposes legacy systems, datawarehouse, forum, document library,
storage and others (without details about others).
Among the authors previously presented, Kerschberg [
41
] also presents some tools to consider
for each layer:
Data mining, data visualization, expert systems, social data mining,
text mining, collaborative filtering, crowdsourcing systems, mash-up,
prediction and idea, markets, syndication systems, trust and
reputation systems.
Business process management systems, configuration management
systems, content management systems, product data management
systems, product lifecycle management systems, ERP, datawarehouse,
decision support systems.
Cloud computing, learning management systems, peer-to-peer
resource sharing, podcasting/videocasting, social media, wiki, audio
conference/video conference, blogs, chat, conversational technologies,
email.
• Presentation: knowledge portal, search services, collaboration and messaging service,
videoconference service, discussion group service and knowledge creation service.
• Knowledge management: data mining services, metadata tagging, ontology and taxonomy,
workflow.
• Knowledge management for information integration services: data warehouse, federation
services, agent services, mediation services, security services.
• Repository: web repository, emails repository, domain repository, text repository and
media repository.
5. Technologies in KM architecture
In this section, the technologies collected are assigned to each layer. The layers considered for
this work, already presented in 3.3 and in Figure 4, are: Presentation, Access and Authentication,
Knowledge Management, Taxonomy and Knowledge Maps, Information Management and
Source / Repository.
First, the technologies of architecture-oriented works are considered [
41, 50, 61
]. The tools
of the collaborative filtering and intelligence layer propose for Xiaohua are considered in KM
layer. The tools of the repository layer are assigned in the layer of the same name. The author
considerer the application layer as "tools enable the integration of information, including implicit
sources of information (for example: people) and explicit sources of knowledge (for example:
databases, transaction storage and data warehouses, etc.), helping to create and share context to
facilitate make judgments, such as brainstorming sessions, problem solving, idea generation,
and strategic planning meetings". For the analysis of this paper, some of the tools are considered
by other authors in the presentation layer and for others in the repository layer, so it is decided
to analyze each tool individually later in this paper.
Kerschberg proposes technologies for the presentation layer, knowledge management and
repository, as well as the N-Layered model considered for the analysis, except "ontology and
taxonomy" which is considered for the specific layer in this work. The fourth layer of Kerschberg
are considered in the detailed analysis presented below.
Cerchione presents a knowledge creation phase on which he presents several tools. These tools
can be included in the Knowledge Management layer, which is responsible for these activities.
Meanwhile, knowledge transfer implies an interaction between knowledge management actors
and therefore the tools reserved for this activity are incorporated in the presentation layer.
Some of the tools proposed by Cordova can also be linked to the proposed layers: intelligence
agents are knowledge management processes (KM Layer), collaboration tools are presentation
tools and proposes an assimilable layer: repositories. Then, the match between [
41, 50, 61
] and
the N-layered are:
• Presentation: knowledge portal, search services, collaboration and messaging service,
videoconference service, discussion group service and knowledge creation service (from
Kerschberg); wiki systems, chats, collaborative forums, blogs, mailing systems (from
Cordova); cloud computing, learning management systems, peer-to-peer resource sharing,
podcasting/videocasting, social media, wiki, audio conference/video conference, blogs,
chat, conversational technologies, email (from Cerchione);
• Knowledge management: agent tools, content personalization, querying, indexing and
metatags (from Xiaohua); data mining services, metadata tagging, workflow (from
Kerschberg); search engines, management reports systems, talent management,
multidimensional cube (from Cordova); data mining, data visualization, expert systems, social data
mining, text mining, collaborative filtering, crowdsourcing systems, mash-up, prediction
and idea, markets, syndication systems, trust and reputation systems (from Cerchione);
• Taxonomy and knowledge Maps: ontology and taxonomy (from Kerschberg).
• Repository: legacy systems, datawarehouse, forum, document library (from Xiaohua),
web repository, emails repository, domain repository, text repository and media repository
(from Kerschberg); knowledge base, question and answer bank, datawarehouse, expert
systems (from Cordova).
The remaining technologies in the diferent papers are presented without assignment to
any layer, so they are analyzed. First, technologies already incorporated in some layers are
eliminated, as are aspects of resources (network, peer-to-peer resource).
Some applications such as Business Process Management Systems, Configuration
Management Systems, Content Management Systems, Product Data Management Systems, Product
Lifecycle Management Systems, ERP, Decision Support Systems (tools presented from
Cerchione), LMS, ECM, Content Management, CRM, platforms (such as e-health, e-gov, etc.)
[
62, 63, 64
], Knowledge Management Application (Open KM, Alfresco, XWiki), RSS, Intelligent
Tutoring System, Business Applications or cloud apps are not feasible to incorporate in a single
layer and must be analyzed in detail in the architecture of each software. They could be called
multilayer software [
65
].
The diferent technologies collected not assigned to any layer are presented in Table 3.
From the resulting list, security services correspond to the Access and Authentication layer.
Technologies involving user interaction are identified and incorporated into the presentation
Video streaming, curricular nets, survey systems, user profiling
systems, visit statistics, customer database, online manuals, procedural
bases, RSS, expert guides.
Skills directory, yellow pages, collaboration tools, digital whiteboards.
Federation services, mediation services, security services.
layer: collaboration tools, digital whiteboards, yellow pages and survey system. For the
repository layer can be identified: online manuals, procedural bases, curricular nets, skills directory,
customer database, data storage, internet, intranet, extranet and expert guides. Meanwhile, the
knowledge management layer can include big data, and text mining.
The user profiling systems or visits statistics can be included in Information Management
layer. The federation services and mediation services are proposed from Kerschberg as an
integration services between the repositories and the KM layer. For the N-layered architecture
at this paper, this role is presented in Information Management layer too.
6. Conclusions
There are several architectures for knowledge management. Most of them are N-Layer
architectures. This paper proposes a 6-layer architecture: Presentation, Access and Authentication,
Knowledge Management, Taxonomy and Knowledge Maps, Information Management and
Source / Repository.
Then, diferent technologies proposed in diferent works are analyzed and assigned to the
corresponding layers, resulting:
• Presentation: knowledge portal, search services, collaboration and messaging service,
videoconference service, discussion group service and knowledge creation service, wiki,
chats, collaborative forums, blogs, mailing systems, cloud computing, Learning
Management Systems, peer-to-peer resource sharing, podcasting, videocasting, social media,
wiki, audio conference, conversational technologies, email, collaboration tools, digital
whiteboards, yellow pages and survey system.
• Access and authentication layer: security services.
• Knowledge management: agent tools, content personalization, querying, indexing and
metatags, data mining services, metadata tagging, workflow, search engines,
Management Reports Systems, talent management, multidimensional cube, Data Mining, data
visualization, expert systems, social data mining, text mining, collaborative filtering,
crowdsourcing systems, mash-up, prediction and idea, markets, syndication systems,
trust and reputation systems, big data, and text mining.
• Taxonomy and Knowledge Maps: ontology and taxonomy.
• Information management: profiling systems, visit statistics, federation services, mediation
services.
• Repository: legacy systems, datawarehouse, forum, document library, web repository,
emails repository, domain repository, text repository and media repository, knowledge
base, question and answer bank, datawarehouse, expert systems, online manuals,
procedural bases, curricular nets, skills directory, customer database, data storage, internet,
intranet, extranet and expert guides.
Some applications are multilayer are not feasible to incorporate in a single layer and must be
analyzed in detail in the architecture of each software.