Онтологічна модель для організації процесу оброблення даних в інформаційних та комунікаційних мережах
The functioning of modern information and communication networks is impossible without data processing. With the emergence of new network services, the amount of information that needs to be processed increases, while the requirements to the data processing quality become more and more stringent. Th...
Збережено в:
| Дата: | 2021 |
|---|---|
| Автори: | , , |
| Формат: | Стаття |
| Мова: | Англійська |
| Опубліковано: |
The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute"
2021
|
| Теми: | |
| Онлайн доступ: | https://journal.iasa.kpi.ua/article/view/236706 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | System research and information technologies |
| Завантажити файл: |  |
Репозитарії
System research and information technologies| _version_ | 1866302739647037440 |
|---|---|
| author | Globa, Larysa Gvozdetska, Nataliia Novogrudska, Rina |
| author_facet | Globa, Larysa Gvozdetska, Nataliia Novogrudska, Rina |
| author_sort | Globa, Larysa |
| baseUrl_str | http://journal.iasa.kpi.ua/oai |
| collection | OJS |
| datestamp_date | 2021-07-13T11:01:37Z |
| description | The functioning of modern information and communication networks is impossible without data processing. With the emergence of new network services, the amount of information that needs to be processed increases, while the requirements to the data processing quality become more and more stringent. Therefore, the problem of designing and maintaining a scalable data processing system with a flexible quality of service management is becoming more and more important for a network operator. Such data processing systems have a complex internal structure with many interrelated parameters, which makes them difficult to analyze, manage, and expand. This study proposes to use an ontological model to store, represent, and manipulate the information in the operator’s data processing system. The ontological model allows to structure and systematize the data of an information processing system, and transparently reflects the relationships between the parameters of the system to simplify its analysis and scaling. The proposed ontology of a data processing system consists of three related subsystems. The paper describes the proposed ontological model and additionally analyzes the sources of information that needs to be processed in the information and communication network. |
| doi_str_mv | 10.20535/SRIT.2308-8893.2021.1.04 |
| first_indexed | 2025-07-17T10:27:14Z |
| format | Article |
| fulltext |
L.S. Globa, N.A. Gvozdetska, R.L. Novogrudska, 2021
Системні дослідження та інформаційні технології, 2021, № 1 47
UDC 004.75
DOI: 10.20535/SRIT.2308-8893.2021.1.04
ONTOLOGICAL MODEL FOR DATA PROCESSING
ORGANIZATION IN INFORMATION
AND COMMUNICATION NETWORKS
L.S. GLOBA, N.A. GVOZDETSKA, R.L. NOVOGRUDSKA
Abstract. The functioning of modern information and communication networks is
impossible without data processing. With the emergence of new network services,
the amount of information that needs to be processed increases, while the require-
ments to the data processing quality become more and more stringent. Therefore, the
problem of designing and maintaining a scalable data processing system with a flex-
ible quality of service management is becoming more and more important for a net-
work operator. Such data processing systems have a complex internal structure with
many interrelated parameters, which makes them difficult to analyze, manage, and
expand. This study proposes to use an ontological model to store, represent, and ma-
nipulate the information in the operator’s data processing system. The ontological
model allows to structure and systematize the data of an information processing sys-
tem, and transparently reflects the relationships between the parameters of the sys-
tem to simplify its analysis and scaling. The proposed ontology of a data processing
system consists of three related subsystems. The paper describes the proposed onto-
logical model and additionally analyzes the sources of information that needs to be
processed in the information and communication network.
Keywords: information and communication network, data processing system, on-
tology, model, network operator, analysis, scaling, class, relations.
INTRODUCTION
Until the last decade, the term “communication network” primarily meant a set of
technical means of communications and facilities designed for routing, switching,
transmission, and/or reception of signals between the terminal equipment [1].
However, in modern communication networks, the operation of technical means
of communications is not possible without software. Moreover, a big part of
hardware functionality is being replaced by software due to the convergence of
information and communications technologies. Communication network becomes
a cyber-physical system, where the physical and software network components
are deeply intertwined. Therefore, the term “information and communications
network” is used instead today to emphasize the fact that the set of information
and communications systems act as a whole within the information processing.
Having an extremely complex internal structure, such a cyber-physical network
can not be easily controlled by the man without using the approaches to intellec-
tual data analysis.
In information and communications network data processing plays an essen-
tial role. To ensure efficient data processing, telecom operators must manage their
data processing system wisely taking into account multiple criteria. Therefore, the
relevant problem for the telecom operator is to represent his data processing sys-
tem in a way to simplify its analysis, maintenance and operation, and to add sup-
L.S. Globa, N.A. Gvozdetska, R.L. Novogrudska
ISSN 1681–6048 System Research & Information Technologies, 2021, № 1 48
port for the system’s scalability. To reach these objectives, the system should be
represented in a formalized way, so that the parameters of the system are explic-
itly exposed, the internal dependencies within the structure are easy to analyze
and the new component may be easily added to support the system’s scalability.
There exist multiple approaches to complex systems’ formalization. In
particular, the system may be represented with the help of the relational model
[2], object-oriented model [2, 3], network-based model [4], ontological model
[5, 6]. The relational model focuses on organizing data in two-dimensional tables.
The advantage of the relational data model lies in the simplicity and convenience
of physical implementation on a computer. The main disadvantages of the
relational model include the lack of standard means of identifying individual
records and the complexity of describing hierarchical and network relationships.
The object-oriented data representation model operates with concepts such as
class and object. Classes define a data structure and represent a set of attributes
(text string, integer, image, etc.). Instances of a class (objects) have a certain
structure and can contain other objects, forming an arbitrary hierarchical
structure. As a rule, systems based on an object-oriented data model are
functional, flexible, but at the same time, more complex. Being hierarchical, these
systems have a limitation: the use case, when a child class has more than one
parent class is not supported here. To tackle this problem, the network-based
model exists. In this model, the classes and relationships are represented in a form
of a graph, which makes this model much more flexible. Ontology is an attempt at
a comprehensive and detailed formalization of a certain area of knowledge using
a conceptual scheme. By the ontological model, data can be represented as a set
of different types of information objects and links between them. The main
advantage of the ontological model is its comprehensiveness. In contrast to the
object-oriented and network-based model, not only the classes and objects are in
focus here, but also the complex semantic relations between them. It helps to
describe the system more exhaustively and explicitly. Because of these
advantages, we have chosen an ontological model to help the operator to simplify
the management and analysis of his data processing system.
In this paper, the usage areas of data processing within the activity of the
information and communication network operator are analyzed. In order to
simplify the data processing organization, an ontology of the operator’s data
processing system is proposed. It formalizes the data processing system and
exposes its parameters, allowing the operator to analyze and manage the data
processing infrastructure with less effort. The paper is structured as follows: in
Section 2 we provide an overview of the related work. Section 3 is dedicated to
the analysis of the main data sources in the operator’s network to reveal the
importance of data processing in modern information and communication
networks. Section 4 describes the proposed ontological model which aims at
simplifying the analysis, management, and scaling of the data processing system.
Section 5 provides an example of simplified data processing system analysis. The
results of the paper are summarized in the Conclusion.
RELATED WORKS
There exist a number of approaches to formalizing a communication network with
the help of an ontological model. The authors of [7] proposed a general ontologi-
cal model that describes a semantic of the domains relevant for the Next Genera-
Ontological model for data processing organization in information and communication networks
Системні дослідження та інформаційні технології, 2021, № 1 49
tion Networks (NGN). Their main idea was to introduce a central point ontology
(core ontology) that defines the main concepts of the mobile domain. Except for
the core ontology, the authors generally described a number of sub-ontologies that
need to be further refined and extended to further domains. Among them is a
communication resources sub-ontology. In this paper we are focused on the data
processing system as a communication resource of the network and thus we aim
at refining this part of the ontological model proposed in [7]. However, from our
point of view the authors did not pay enough attention to the quality of service
assessment in the communication network, although this aspect itself contains a
large number of concepts and relations that should be considered and formalized.
In [8] an approach to ontology modelling for telecommunications service
domain was proposed. Within their telecommunications service domain ontology
(TSDO), the authors distinguish the Terminal Capability Ontology, Network
Ontology, Service Role Ontology, Service Category Ontology, Charging
Ontology and Service Quality Ontology. From our point of view, the main
advantage of the proposed concept is taking into account the quality of service
parameters as a separate full-fledged ontology. Meanwhile, the authors paid
attention to formalizing the communication resource domain as well. In our work
we would like to elaborate on this concept and describe the network resources in a
relation with its quality parameters in particular. Since the service quality is
influenced not only by the network resources but also by the various workloads
which have to be processed in the network, in contrast to [8] we would like to pay
separate attention to the semantic interoperability of all three domains: network
resources, service quality, and workload.
According to the 5G whitepaper [9], in new networks along with traditional
Quality of Service parameters such as the packet error rate, transmission latency,
and data rate, the new parameters such as network energy efficiency are becoming
more and more important. Energy efficiency of data processing in general is a
highly relevant topic. According to [10] the amount of power consumed by the
data processing facilities around the world comprises near 2% of all electrical
power produced worldwide. In order to deal with this problem a number of
hardware and software energy efficient approaches to data processing were
proposed. Among them are the virtual machines consolidation [11], energy
efficient scheduling [12, 13], resource scaling [14]. These approaches have
already become standard for the distributed data processing facilities, however,
they did not get enough attention in the communication network domain. Of
course, energy efficient approaches are used there, but from our point of view,
they must be included in the general formalized data processing architecture. In
this paper, we try to approach this problem with the proposed ontological model
of the data processing system in communications.
DATA PROCESSING RESOURCES IN INFORMATION AND
COMMUNICATION NETWORKS
According to the authors of [15], information and communication network re-
sources are divided into information, data processing and storage resources, soft-
ware. Information resources are information and knowledge transmitted through
the information and communication network. Data processing and storage re-
sources are the performance of processors and the amount of memory of com-
L.S. Globa, N.A. Gvozdetska, R.L. Novogrudska
ISSN 1681–6048 System Research & Information Technologies, 2021, № 1 50
puters running on the network, as well as the time during which they are used.
Software resources include network operating systems, server software, work-
station software, application software, traffic analyzers, network controls, and
more. Communication resources are resources that are involved in the transporta-
tion and redistribution of information flows in a network. It means that the data
processing and storage resources have already become an essential part of modern
communication networks.
Alongside with the access and core network, data centers (representing the
data processing and storage resources) are becoming the key components of the
infrastructure of the communication network operator [16]. Let us briefly
overview these key components of the network. Access network is connected to
the end (terminal) nodes — equipment installed by users of the network. For
example, in the case of building an operator network to provide Internet access
services, the end nodes may be subscribers’ computers or subscriber routers. The
main purpose of the access network is to concentrate the information flows
coming through numerous communication channels from user equipment. The
core network combines individual access networks, performing the functions of
traffic transit between them through high-speed channels. Data centers and
service management centers are network resources on the basis of which customer
service is provided. Such centers can store information of two types:
user information, i.e., information that is of direct interest to end users of
the network (information resources);
service information that helps provide services to users.
Examples of the first-type information resources are web-portals, which
contain a variety of reference and news information, information from e-shops,
etc. Resources of the second type are various systems of authentication and
authorization allowing the operator to check the rights of users for receiving the
services; billing systems, which help to manage charges for services in commercial
networks; databases of user credentials that store usernames and passwords, as
well as lists of services to which each user is subscribed, etc. The second type of
resources should also include a centralized network management system.
A prominent example of an information and communication network in
which information and communication technologies work as a single indivisible
whole is the 5G network. Let us briefly consider the basic principles of this
network design in order to reveal the additional purposes of data processing in 5G
networks. To do this, let us analyze the white paper provided by the European
organization 5G PPP (public-private partnership in the field of 5G
infrastructure — a joint initiative between the European Commission and the
European ICT industry (ICT manufacturers, communications operators, service
providers, SMEs and research institutions). According to the documentation [9],
the key paradigm of the 5th generation mobile network is its programmability.
Programmability ensures flexible network adaptation at various levels, including
infrastructure, network functions, services, and applications. In particular
programmability in the data plane, transport network (core network) and access
network are distinguished. In 5G programmability is primarily inspired by two
technologies: SDN (Software Defined Network) and NFV (Network Function
Virtualization) technologies.
SDN is an approach to network design, implementation, and management
that separates network management (control plane) and traffic management process
(data plane). This separation greatly simplifies network administration and
Ontological model for data processing organization in information and communication networks
Системні дослідження та інформаційні технології, 2021, № 1 51
management, as the control plane processes only information related to the logical
topology of the network. The data plane instead organizes network traffic accord-
ing to the configuration set in the control plane. Unlike conventional IP networks,
whose functions are decentralized, SDN follows a centralized approach [17].
The main idea of NFV technology is to replace specialized network
equipment (e.g., L2 switches, routers, NAT devices (Network Address
Translation), firewalls, etc.) with software — virtualized network functions —
consolidated on general-purpose hardware (commodity servers) [18].
Thus, to enable these technologies communication operators must maintain
the data processing infrastructure as well.
Another purpose of maintaining the data processing infrastructure for the
telecom operator is a Big Data analysis. Such an analysis helps operators to
improve the technical and economic parameters of the network, enable the
personalization of telecom services, and ensure more efficient allocation of funds.
Data processing infrastructure is used in particular for the analysis of the
subscribers’ activity, traffic changes, long-term network characteristics, etc. The
operators report the positive impact of the Big Data analysis included in their
operational workflow [19, 20].
Summarizing this analysis, we would like to highlight 3 main areas of use of
data processing by a telecom operator:
1) ensuring the functioning of SDN and NFV technologies;
2) implementation of such necessary network functions as authorization and
authentication of users, billing service, etc.;
3) analysis of Big Data in the field of communications.
Thus, the effective design, construction and operation of data processing
infrastructure is an important problem for operators of modern communications
networks (including 5G networks).
ONTOLOGICAL MODEL OF A DATA PROCESSING SYSTEM
IN INFORMATION AND COMMUNICATION NETWORKS
Computational ontologies are the means to formally model the structure of a sys-
tem, i.e., the relevant entities and relations that emerge from its observation, and
which are useful to our purposes [5]. Ontologies (or ontological models) help to
formalize the structure of a system in order to simplify its management, improve
its design, automatize system’s operation, etc. In this paper we introduce the onto-
logical model for the data processing system of the telecom operator to simplify
its management. This simplification is a consequence of considering the system as
a formalized structure with explicitly exposed parameters.
Being a structured representation of the information in some subject area,
every ontology is based on the raw data, stored in some kind of informational da-
tabase (Fig. 1). Examples of such raw data are the subscribers’ records in a billing
system, statistics of the daily workload in a data processing system, nominal pa-
rameters of the data processing equipment, etc. The ontological model transforms
this raw data into knowledge. Modeling the data processing system as an ontol-
ogy, we may distinguish 3 separate structural parts of the system. These parts cor-
respond to the 3 subcomponents of the ontological model respectively:
1) the ontology of the processing system components (servers, processing
software, etc.);
L.S. Globa, N.A. Gvozdetska, R.L. Novogrudska
ISSN 1681–6048 System Research & Information Technologies, 2021, № 1 52
2) the ontology of the workloads (the input workload sources and parame-
ters, etc.);
3) the ontology of the system quality assessment criteria (performance, en-
ergy efficiency, etc).
Dividing general ontology of Data Processing System in to such correlated
components allows to describe in detailed way each subprocess that is performed
be network operator while analyzing and managing data processing infrastructure
with less effort.
Let us explain these subsystems. Data processing equipment and software
represent the ontology of processing system components. The ontology of these
components reflects the physical structure of a system itself. The ontology of the
workloads encompasses the input workload and parameters of the available data
processing equipment. The input workload comes from the aforementioned
sources (we consider either Big Data tasks, subscribers’ data or NFV processing
tasks to be the sources here). Computational job is a unit of workload. Each job is
characterized by its requirements which depend on the type of the workload
source. The processing system consists of server clusters, consisting of N physical
nodes (servers) respectively. Each server has an amount of computational
resources (processing cores, RAM, etc.) and operates with specialized data
processing software, aimed to manage the resources of the processing system,
schedule the computational jobs, etc.
The output parameters are those criteria that are evaluated for the system.
For the efficient data processing, telecom operator needs to design his data proc-
essing systems in a way to:
fulfill the Quality of service requirements depending on the workload type
(e.g., the allowed probability of a job loss for the internal Big Data processing
tasks may be higher than the probability of loss for the billing service task, since
it presumes direct cooperation with the subscriber);
ensure sufficient performance of the data processing system (with
“performance” here we mean the amount of processed data per time unit. I.e., the
bigger performance of the data processing system is, the higher processing
throughput is achieved);
increase the energy efficiency of the data processing system. Since it
greatly affects the operator’s OPEX (Operating expense), so it is important to
consider this criterion not only during system design (to purchase more energy
efficient equipment), but also during operation.
Fig. 1. General structure of the proposed ontological model of the data processing system
Ontological model for data processing organization in information and communication networks
Системні дослідження та інформаційні технології, 2021, № 1 53
These parameters are described by the ontology of the system quality as-
sessment criteria.
Formally, the ontology may be specified as a set [6]:
},,,,,{ DFTRACO ,
where
1) C is the set of classes that describes the notions of a subject domain;
2) A is the set of attributes that describes the features of notions and rela-
tions;
3) R is the set of relations specified for classes:
},,,,{ CDnIAAS RRRRR
where ASR is the associative relation:
},{)(),()({)( strRMOCOCOR ASjiAS
where M is a type of relation meaning,
IAR is the relation “is–are”, also known as a “part–whole” relation:
)()()( OCOCOR mkIA ,
nR is the relation of inheritance:
)(,)(,)( OAraOAraOR
km CiiCiin ,
CDR is the relation “class–data”:
)()()( ODOCOR ijCD ;
4) T is the set of standard types of attribute values;
5) F is the set of limits for values of attribute notions and relations;
6) D is the set of instances for a particular class.
The proposed ontology of the telecom operator’s data processing system is
represented in Fig. 2. The formal description of the ontology may be found below.
Proposed ontology is formally described as follows:
Set of ontology classes: },...,,{ 2521 CCCC
1C — Operator’s data processing system. This class describes the concept of
an operator’s data processing system as a physical entity.
2C — Source of the workload. This class includes concepts related to the
input workload for the data processing system. As discussed in Section 1, possible
sources (at least those considered in this paper) are.
3C — Operator’s Big Data.
4C — Subscribers’ data (e.g., credentials, data about the usage of services,
billing information, etc.).
5C — NFV processing task. Within the NFV concept, the tasks usually per-
formed by the specialized hardware (e.g., traffic routing, NAT, etc.) are per-
formed on the commodity servers. Within this ontology we call these tasks gener-
ally “NFV processing tasks”, however this entity may be refined with respect to
the concrete virtualized functions in the considered network.
L.S. Globa, N.A. Gvozdetska, R.L. Novogrudska
ISSN 1681–6048 System Research & Information Technologies, 2021, № 1 54
F
ig
. 2
. T
he
p
ro
po
se
d
on
to
lo
gy
o
f
th
e
ne
tw
or
k
op
er
at
or
’s
d
at
a
pr
oc
es
si
ng
s
ys
te
m
Ontological model for data processing organization in information and communication networks
Системні дослідження та інформаційні технології, 2021, № 1 55
These classes are connected with their parent class 2C with the help of the
relation of inheritance.
6C — Server cluster. A class describes physical objects that are part of a data
processing system. The attributes of the server cluster are the number of its nodes,
load and cluster model as a queuing system: ),( 666 DAC .
7C — Server. This class describes the physical objects that are part of the
server cluster. The attributes of the server are its power consumption function and
computational resources: ),( 777 DAC .
8C — Computational resource. This class includes concepts that describe
the physical computing resources of the server. The attribute of the computing
resource is its volume: ),( 888 DAC Inherited types of the computational
resources are:
9C — RAM (Random Access Memory). This class reflects the concept of
the physical resource of the server’s RAM. The class has an inherited “volume”
attribute that displays the amount of available server RAM: ),( 999 DAC .
10C — Processing core. The class describes the concept of the server’s
physical processor resource. It has an inherited “volume” attribute that displays
the number of server processing cores: ),( 101010 DAC .
11C — Data storage. This class displays the concept of a physical resource -
a data storage device (hard disk, SSD, etc.). The class has an inherited attribute
“volume”, which reflects the available amount of server storage: ),( 111111 DAC .
These classes are connected with their parent class 8C with the help of the
relation of inheritance .
8nR
12C — Workload. This class is responsible for describing the abstract
concept of input workload of a data processing system. The workload is generated
by the sources of the workload )( 2C . It is characterized by the attribute
“statistical model”, which corresponds to the concept of the statistical curve of the
input load: ),( 121212 DAC .
13C — Computational job. This class describes the concept of a
computational job as a unit of input workload in the system. The class has a
“volume” attribute that expresses the amount of computations required to process
the job (for example, in the number of elementary operations): ),( 131313 DAC .
14C — Requirement. This class describes the requirements of computational
jobs for the physical resources of the system. The requirement has a job ID as an
attribute to bind a specific requirement to the job that owns it: ),( 141414 DAC .
The inherited requirements are:
15C — Maximum allowed processing time. This is the time after which the
job is removed from the processing, even if the processing was not completed
successfully. The class inherits the “job ID” attribute to bind a specific job
requirement to its job: ),( 151515 DAC .
16C — Minimum memory (capacity). The class describes one of the
requirements of a computational job, which reflects the minimum amount of
RAM of a server, at which a job can still be processed on this server. The class
inherits the “job ID” attribute to bind a specific requirement to the job that owns
it: ),( 161616 DAC .
L.S. Globa, N.A. Gvozdetska, R.L. Novogrudska
ISSN 1681–6048 System Research & Information Technologies, 2021, № 1 56
17C — Number of cores. The class describes one of the computational job’s
requirements, which reflects the minimum number of free server processor cores
that will be allocated for processing this job. The class inherits the “job ID”
attribute to bind a specific job requirement to its job: ),( 171717 DAC .
18C — Minimum available storage. The class describes one of the
computational requirements that reflects the minimum amount of free server
storage to be used during processing. The class inherits the “job ID” attribute to
bind a specific job requirement to the corresponding job: ),( 171717 DAC .
These classes are connected with their parent class 14C with the help of the
relation of inheritance .
14nR
19C — Quality of Service (QoS) requirements. The class describes the
requirements to the data processing that correspond to a specific type of service
(data being processed). In our interpretation the class has an attribute “maximum
allowed job loss probability”, which limits the probability of losing the job when
processing by the system: ),( 181818 DAC . In general, any other attributes agreed
within the QoS requirements may be shown here.
20C — Energy efficiency. Describes the concept of energy efficiency of a
processing system in general and each physical server in particular as the amount
of electrical power consumed to perform a unit of work.
21C — Performance. Describes the concept of performance of a computer
system in general and each physical server in particular as the amount of work
performed per unit time.
22C — Data processing software. Displays the concept of software used in a
distributed computing system to distribute and process the workload, as well as
control the state of the system as a whole. In our interpretation it encompasses the
following types of software:
23C — Scheduling software. This class represents the software that is used
to distribute the computational jobs between the available hardware (servers).
This process is also known as jobs’ scheduling.
24C — Scaling software. This type of software is responsible for managing
the quantity of available hardware in the system. It is in particular important for
energy efficiency: fewer resources may be kept available in case of the underload
of the system to save some power.
25C — Consolidation software. This software manages the consolidation
process of the virtual machines.
If other types of software are used by the operator, they may be added as the
separate classes of the ontology as well.
These classes are connected with their parent class 22C with the help of the
relation of inheritance .
22nR
Associative relations: }{ jiAS XCCR
“produces” — displays the relationship between the “Source of the work-
load” class and the “Workload” and shows the process of producing the workload
by various sources;
“influences” — reproduces the relation between the “Source of the
workload” and “QoS requirements” and expresses, that different types of
workload sources have different QoS requirements;
Ontological model for data processing organization in information and communication networks
Системні дослідження та інформаційні технології, 2021, № 1 57
“influences fulfilment” — connects the classes “Workload” and “QoS
requirements” and shows that the amount of workload influences the fulfillment of the
QoS requirements (e.g., it is more difficult to fulfill the requirements in busy hours);
“provides fulfillment” — this relationship describes the relationship between
the “Operator’s data processing system” and “QoS requirements”. The system
must operate in such a way as to ensure compliance with the QoS requirements;
“defines” — this relationship describes the relationship between the class
“Computational resource” and “Performance”. The essence of this relationship is
to reflect the impact of the quantity and quality of server computing resources on
its performance. This relation connects the classes “Requirements” and “Server”
as well to show the fact that the jobs’ requirements define the choice of the
hardware, on which the job may be processed;
“has” — a connection that shows the logical affiliation of one class to
another. The classes “Server” — ”Computing resource”, “Server” — “Data
processing software”, “Computational job” — “Requirement” have this relation;
“provides processing” — this relation connects the classes “Data processing
software” and “Computational job” showing, that the data processing software
operates with the jobs in order to let them be processed;
“manages” — the relation shows, that the “Computational resources” are
managed (scaled, distributed, etc.) with the help of the “Data processing
software”;
“evaluated for” — shows the relationship between the parameters of the data
processing system (Energy Efficiency and Performance) and the Operator’s data
processing system itself.
“Part-whole” relations: )()()( OCOCOR mkIA
“Part-whole” relations are defined between the classes “Operator’s data
processing system” and “Server cluster”, “Server cluster” and “Server”,
“Workload” and “Computational job” to show that one entity is a part of another
one.
The described ontology formalizes the data processing system and simplifies
the analysis and management of such a system. The parameters of the system are
explicitly exposed and the operator is able to see the relations between them.
AN EXAMPLE OF A SIMPLIFIED DATA PROCESSING ORGANIZATION
WITH THE HELP OF THE ONTOLOGICAL MODEL
Let us consider an example of simplified data processing organization with the
help of ontology. In the related research [21], we consider the problem of energy
efficient data processing which is a very important topic nowadays. The problem
is to ensure a minimal power consumption of a data processing system without
losing the processing performance and ensuring the fulfilment of the QoS re-
quirements. This is a complicated task which requires having an overview on a
system as a whole, and taking into account multiple influencing parameters. Due
to the task’s specificity, it is infeasible to analyze the separate parts of the system,
since they cooperate solving the processing tasks and act together as a single dis-
tributed data processing system.
The proposed ontology explicitly shows the complex semantical dependen-
cies between the input workload parameters of the system and assessment criteria.
The operator sees that the fulfilment of the QoS requirements is directly influ-
enced by the system performance and indirectly influenced by the computational
L.S. Globa, N.A. Gvozdetska, R.L. Novogrudska
ISSN 1681–6048 System Research & Information Technologies, 2021, № 1 58
resources of the system. Thus, in order to fulfil the QoS requirements he should
increase the volume of the computational resources. However, he also sees that
the resources are managed by the data processing software and so, instead of
changing the resources (which may be costly) the operator may focus on the soft-
ware tuning in order to improve the resource management.
Considering the energy efficiency criterion, the operator sees that it is
defined for each separate processing node (server) in the system. And the decision
regarding the server to be chosen is taken based on the requirements of the input
jobs. Therefore, a design decision based on this analysis would be to pay attention
to the jobs’ requirements analysis to ensure a more thoughtful choice of the
processing server.
Based on a conducted analysis, a comprehensive energy efficient approach
to workload processing was proposed in [21]. This approach takes into account
individual power consumption characteristics of computing nodes, deals with dy-
namic workload deviations, and ensures meeting requirements to the service qual-
ity combining energy efficient scheduling and horizontal scaling. The results of
the approach are largely due to the ontological model, which helped to identify
and link together all the complex semantic dependencies of the system. All the
details regarding the approach and its evaluation may be found in [21]. The main
point that we would emphasize here is that due to the used ontological model, the
complex dependencies between the parameters and the assessment criteria of the
system were easily embraced and the formalized system representation was used
as an input for the automatic system optimization software. The ontological mod-
el is designed and refined once for the whole system and helps to analyze it in the
future due to the visualization and formalization.
CONCLUSION
In this paper we analyzed how the data processing is involved in the range of ac-
tivities of the modern information and communication network operator. We de-
fined that the main directions of data processing in modern information and
communication networks are related to the NFV and SDN applications, traditional
subscribers’ management functions (authorization and authentication of users,
billing service, etc.) and Big Data processing. In order to simplify the organiza-
tion of this data processing, an ontological model of the data processing system in
communications was proposed. This model formalizes the data processing system
exposing its parameters and visualizing the relations between them. It simplifies
the analysis of the system for the network operator and enables partial or full au-
tomatization of the system analysis and management in future. An example of an
energy efficient data processing problem was considered to show how an onto-
logical model simplifies the analysis and optimization of a complex data process-
ing system.
The proposed ontology assumes the possibility of expansion and addition.
For example, processing security can be considered as another important criterion
for the quality of data processing (especially for the modern network services
such as connected vehicles). This parameter and corresponding additions to the
ontology should be considered as a related future work.
REFERENCES
1. The law of Ukraine About telecommunications [Online]. Available:
https://zakon.rada.gov.ua/laws/show/1280-15#Text). Accessed on: 11 Oct. 2020.
Ontological model for data processing organization in information and communication networks
Системні дослідження та інформаційні технології, 2021, № 1 59
2. Jan Harrington, The Relational Data Model, 2009. doi: 10.1016/B978-0-12-374730-
3.00005-X.
3. Antonio Badia, Relational, Object-Oriented and Object-Relational Data Models,
2006. doi: 10.4018/9781591405603.ch088.
4. John Mariani and Tom Rodden, A Model for Schema Evolution in Object-Oriented
Database Systems, 1994.
5. Nicola Guarino, Daniel Oberle, and Steffen Staab, What Is an Ontology?, 2009. doi:
10.1007/978-3-540-92673-3_0.
6. L.S. Globa, R.L. Novogrudska, and A.V. Koval, “Ontology Model of Telecom Op-
erator Big Data”, Proceedings of IEEE International Black Sea Conference on
Communication and Netwoorking (BlackSeaCom), 1-5, 2018. doi:10.1109/ Black-
SeaCom. 2018.8433710
7. C. Villalonga et al., “Mobile Ontology: Towards a Standardized Semantic Model for
the Mobile Domain”, Proceedings of the 1st International Workshop on Telecom
Service Oriented Architectures (TSOA 2007) at the 5th International Conference on
Service-Oriented Computing, Vienna, Austria, September 17, 2007, pp. 248–257.
8. Qiao Xiuquan, Li Xiaofeng, and Chen Junliang, Telecommunications Service Do-
main Ontology: Semantic Interoperation Foundation of Intelligent Integrated Ser-
vices, 2012. doi: 10.5772/36794.
9. View on 5G Architecture: 5G PPP Architecture Working Group [Online]. Available:
https://5g-ppp.eu/wp-content/uploads/2019/07/5G-PPP-5G-Architecture-White-
Paper_v3.0_PublicConsultation.pdf. Accessed on: 20 Sept. 2020.
10. Data Center Market and Technology Trends Power Electronics presentation
[Online]. Available: https://www.slideshare.net/Yole_Developpement/data-center-
market-and-technology-trends-power-electronics-presentation-held-at-apec-2016-
from-yole-dveloppement. Accessed on: 18 Oct. 2020.
11. Qiu Yeliang, Jiang Congfeng, Wang Yumei, Ou Dongyang, Li Youhuizi, and Wan
Jian, Energy Aware Virtual Machine Scheduling in Data Centers. Energies, 2019.
doi: 12. 646. 10.3390/en12040646.
12. F. Armenta-Cano et al., “Min_c: Heterogeneous Concentration Policy for Power Aware
Scheduling. Trudy ISP RAN”, Proc. ISP RAS, vol. 27, issue 6, 2015, pp. 355–380.
13. Hosseinimotlagh Seyedmehdi, Khunjush Farshad, and Hosseinimotlagh Seyedma-
hyar, A Cooperative Two-Tier Energy-Aware Scheduling for Real-Time Tasks in
Computing Clouds, 2014, pp. 178–182. doi: 10.1109/PDP.2014.91.
14. Mohammad Aldossary and Karim Djemame, Performance and Energy-Based Cost
Prediction of Virtual Machines Auto-Scaling in Clouds, 2018.
15. P.P. Vorobienko, L.A. Nikityuk, and P.I. Reznichenko, Communication and infor-
mation networks, Approved by the Ministry of Education and Science of Ukraine as a
textbook for students of higher education institutions. Kyiv, 2010.
16. Talapko Domagoj, Telecom datacenter power infrastructure availability comparison
of DC and AC UPS, 2012, pp. 1–5. doi: 10.1109/INTLEC.2012.6374509.
17. Benzekki Kamal, El Fergougui Abdeslam, El Belrhiti El Alaoui Abdelbaki, Soft-
ware-defined networking (SDN): A survey. Security and Communication Networks,
2017. doi: 9. 10.1002/sec.1737.
18. Zhang Tianzhu, NFV Platform Design: A Survey, 2020.
19. Big Data for business. Big Data from Kyivstar [Online]. Available:
https://kyivstar.ua/uk/business/products/big-data. Accessed on: 13 Oct. 2020.
20. “Cellular operators know everything about their subscribers”, Vedomosti. Accessed
on: 13 Oct. 2020. [Online]. Available: https://www.vedomosti.ru/technology/ arti-
cles/2015/05/26/593579 -sotovie-operatori-znayut-vse-o-svoih-abonentah
21. L. Globa and N. Gvozdetska, “Comprehensive Energy Efficient Approach to Work-
load Processing in Distributed Computing Environment”, 2020 IEEE International
Black Sea Conference on Communications and Networking (BlackSeaCom), Odessa,
Ukraine, 2020, pp. 1–6. doi: 10.1109/BlackSeaCom48709.2020.9235010.
Received: 07.12.2020
L.S. Globa, N.A. Gvozdetska, R.L. Novogrudska
ISSN 1681–6048 System Research & Information Technologies, 2021, № 1 60
INFORMATION ON THE ARTICLE
Larysa S. Globa, ORCID: 0000-0003-3231-3012, National Technical University of
Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Ukraine, e-mail: lgloba@its.kpi.ua
Nataliia A. Gvozdetska, ORCID: 0000-0001-6549-0459, National Technical University
of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Ukraine, e-mail:
n.gvozdetska@gmail.com
Rina L. Novogrudska, ORCID: 0000-0002-0533-5817, National Technical University of
Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Ukraine, e-mail: rinan@ukr.net
ОНТОЛОГІЧНА МОДЕЛЬ ДЛЯ ОРГАНІЗАЦІЇ ПРОЦЕСУ ОБРОБЛЕННЯ
ДАНИХ В ІНФОРМАЦІЙНИХ ТА КОМУНІКАЦІЙНИХ МЕРЕЖАХ /
Л.С. Глоба, Н.А. Гвоздецька, Р.Л. Новогрудська
Аннотація. Функціонування сучасних інформаційно-телекомунікаційних ме-
реж неможливе без оброблення даних. З появою нових мережевих послуг
кількість інформації, що потребує оброблення, зростає, при цьому ставляться
дедалі жорсткіші вимоги до якості оброблення даних. Тому для оператора ме-
режі дедалі більшої актуальності набуває проблема побудови та підтримання
системи оброблення даних з можливістю гнучкого керування якістю послуг та
масштабування. Такі системи оброблення даних мають комплексну внутрішню
структуру з багатьма взаємопов’язаними параметрами, що ускладнює їх аналіз,
керування та розширення. Запропоновано використовувати онтологічну мо-
дель для зберігання, подання та маніпулювання інформацією в системі оброб-
лення даних оператора. Онтологічна модель дозволяє структурувати та систе-
матизувати дані системи оброблення інформації і прозоро відображати
взаємозв’язки між параметрами системи для спрощення її аналізу та масшта-
бування. Запропонована онтологія системи оброблення даних складається з
трьох зв’язаних підсистем. Наведено опис запропонованої онтологічної моделі
та додатково проаналізовано джерела інформації, яка потребує оброблення, в
інформаційно-телекомунікаційній мережі.
Ключові слова: інформаційно-телекомунікаційна мережа, система оброблен-
ня даних, онтологія, модель, оператор мережі, аналіз, масштабування, клас,
відношення.
ОНТОЛОГИЧЕСКАЯ МОДЕЛЬ ДЛЯ ОРГАНИЗАЦИИ ПРОЦЕССА
ОБРАБОТКИ ДАННЫХ В ИНФОРМАЦИОННЫХ И КОММУНИКАЦИОННЫХ
СЕТЯХ / Л.С. Глоба, Н.А. Гвоздецкая, Р.Л. Новогрудская
Аннотация. Функционирование современных информационно-телекомму-
никационных сетей невозможно без обработки данных. С появлением новых
сетевых услуг количество информации, которая нуждается в обработке, воз-
растает, при этом выдвигаются все более жесткие требования к качеству обра-
ботки данных. Поэтому для оператора сети всё большую актуальность приоб-
ретает проблема построения и поддержки системы обработки данных
с возможностью гибкого управления качеством услуг и масштабирования. Та-
кие системы обработки данных имеют комплексную внутреннюю структуру со
многими взаимосвязанными параметрами, что затрудняет их анализ, управле-
ние и расширение. Предложено использовать онтологическую модель для хра-
нения, представления и манипулирования информацией в системе обработки
данных оператора. Онтологическая модель позволяет структурировать и сис-
тематизировать данные системы обработки информации и прозрачно отражать
взаимосвязи между параметрами системы для упрощения её анализа и мас-
штабирования. Предложенная онтология системы обработки данных состоит
из трех связанных подсистем. Приведено описание предложенной онтологиче-
ской модели и дополнительно проанализированы источники информации, ко-
торая нуждается в обработке, в информационно-телекоммуникационной сети.
Ключевые слова: информационно-телекоммуникационная сеть, система об-
работки данных, онтология, модель, оператор сети, анализ, масштабирование,
класс, отношения.
|
| id | journaliasakpiua-article-236706 |
| institution | System research and information technologies |
| keywords_txt_mv | keywords |
| language | English |
| last_indexed | 2025-07-17T10:27:14Z |
| publishDate | 2021 |
| publisher | The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" |
| record_format | ojs |
| resource_txt_mv | journaliasakpiua/dc/4ebcec21c6a808983d358096ca84ffdc.pdf |
| spelling | journaliasakpiua-article-2367062021-07-13T11:01:37Z Ontological model for data processing organization in information and communication networks Онтологическая модель для организации процесса обработки данных в информационных и коммуникационных сетях Онтологічна модель для організації процесу оброблення даних в інформаційних та комунікаційних мережах Globa, Larysa Gvozdetska, Nataliia Novogrudska, Rina информационно-телекоммуникационная сеть система обработки данных онтология модель оператор сети анализ масштабирование класс отношения information and communication network data processing system ontology model network operator analysis scaling class інформаційно-телекомунікаційна мережа система оброблення даних онтологія модель оператор мережі аналіз масштабування клас відношення The functioning of modern information and communication networks is impossible without data processing. With the emergence of new network services, the amount of information that needs to be processed increases, while the requirements to the data processing quality become more and more stringent. Therefore, the problem of designing and maintaining a scalable data processing system with a flexible quality of service management is becoming more and more important for a network operator. Such data processing systems have a complex internal structure with many interrelated parameters, which makes them difficult to analyze, manage, and expand. This study proposes to use an ontological model to store, represent, and manipulate the information in the operator’s data processing system. The ontological model allows to structure and systematize the data of an information processing system, and transparently reflects the relationships between the parameters of the system to simplify its analysis and scaling. The proposed ontology of a data processing system consists of three related subsystems. The paper describes the proposed ontological model and additionally analyzes the sources of information that needs to be processed in the information and communication network. Функционирование современных информационно-телекоммуникационных сетей невозможно без обработки данных. С появлением новых сетевых услуг количество информации, которая нуждается в обработке, возрастает, при этом выдвигаются все более жесткие требования к качеству обработки данных. Поэтому для оператора сети всё большую актуальность приобретает проблема построения и поддержки системы обработки данных с возможностью гибкого управления качеством услуг и масштабирования. Такие системы обработки данных имеют комплексную внутреннюю структуру со многими взаимосвязанными параметрами, что затрудняет их анализ, управление и расширение. Предложено использовать онтологическую модель для хранения, представления и манипулирования информацией в системе обработки данных оператора. Онтологическая модель позволяет структурировать и систематизировать данные системы обработки информации и прозрачно отражать взаимосвязи между параметрами системы для упрощения её анализа и масштабирования. Предложенная онтология системы обработки данных состоит из трех связанных подсистем. Приведено описание предложенной онтологической модели и дополнительно проанализированы источники информации, которая нуждается в обработке, в информационно-телекоммуникационной сети. Функціонування сучасних інформаційно-телекомунікаційних мереж неможливе без оброблення даних. З появою нових мережевих послуг кількість інформації, що потребує оброблення, зростає, при цьому ставляться дедалі жорсткіші вимоги до якості оброблення даних. Тому для оператора мережі дедалі більшої актуальності набуває проблема побудови та підтримання системи оброблення даних з можливістю гнучкого керування якістю послуг та масштабування. Такі системи оброблення даних мають комплексну внутрішню структуру з багатьма взаємопов’язаними параметрами, що ускладнює їх аналіз, керування та розширення. Запропоновано використовувати онтологічну модель для зберігання, подання та маніпулювання інформацією в системі оброблення даних оператора. Онтологічна модель дозволяє структурувати та систематизувати дані системи оброблення інформації і прозоро відображати взаємозв’язки між параметрами системи для спрощення її аналізу та масштабування. Запропонована онтологія системи оброблення даних складається з трьох зв’язаних підсистем. Наведено опис запропонованої онтологічної моделі та додатково проаналізовано джерела інформації, яка потребує оброблення, в інформаційно-телекомунікаційній мережі. The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" 2021-07-13 Article Article application/pdf https://journal.iasa.kpi.ua/article/view/236706 10.20535/SRIT.2308-8893.2021.1.04 System research and information technologies; No. 1 (2021); 47-60 Системные исследования и информационные технологии; № 1 (2021); 47-60 Системні дослідження та інформаційні технології; № 1 (2021); 47-60 2308-8893 1681-6048 en https://journal.iasa.kpi.ua/article/view/236706/235248 |
| spellingShingle | інформаційно-телекомунікаційна мережа система оброблення даних онтологія модель оператор мережі аналіз масштабування клас відношення Globa, Larysa Gvozdetska, Nataliia Novogrudska, Rina Онтологічна модель для організації процесу оброблення даних в інформаційних та комунікаційних мережах |
| title | Онтологічна модель для організації процесу оброблення даних в інформаційних та комунікаційних мережах |
| title_alt | Ontological model for data processing organization in information and communication networks Онтологическая модель для организации процесса обработки данных в информационных и коммуникационных сетях |
| title_full | Онтологічна модель для організації процесу оброблення даних в інформаційних та комунікаційних мережах |
| title_fullStr | Онтологічна модель для організації процесу оброблення даних в інформаційних та комунікаційних мережах |
| title_full_unstemmed | Онтологічна модель для організації процесу оброблення даних в інформаційних та комунікаційних мережах |
| title_short | Онтологічна модель для організації процесу оброблення даних в інформаційних та комунікаційних мережах |
| title_sort | онтологічна модель для організації процесу оброблення даних в інформаційних та комунікаційних мережах |
| topic | інформаційно-телекомунікаційна мережа система оброблення даних онтологія модель оператор мережі аналіз масштабування клас відношення |
| topic_facet | информационно-телекоммуникационная сеть система обработки данных онтология модель оператор сети анализ масштабирование класс отношения information and communication network data processing system ontology model network operator analysis scaling class інформаційно-телекомунікаційна мережа система оброблення даних онтологія модель оператор мережі аналіз масштабування клас відношення |
| url | https://journal.iasa.kpi.ua/article/view/236706 |
| work_keys_str_mv | AT globalarysa ontologicalmodelfordataprocessingorganizationininformationandcommunicationnetworks AT gvozdetskanataliia ontologicalmodelfordataprocessingorganizationininformationandcommunicationnetworks AT novogrudskarina ontologicalmodelfordataprocessingorganizationininformationandcommunicationnetworks AT globalarysa ontologičeskaâmodelʹdlâorganizaciiprocessaobrabotkidannyhvinformacionnyhikommunikacionnyhsetâh AT gvozdetskanataliia ontologičeskaâmodelʹdlâorganizaciiprocessaobrabotkidannyhvinformacionnyhikommunikacionnyhsetâh AT novogrudskarina ontologičeskaâmodelʹdlâorganizaciiprocessaobrabotkidannyhvinformacionnyhikommunikacionnyhsetâh AT globalarysa ontologíčnamodelʹdlâorganízacííprocesuobroblennâdanihvínformacíjnihtakomuníkacíjnihmerežah AT gvozdetskanataliia ontologíčnamodelʹdlâorganízacííprocesuobroblennâdanihvínformacíjnihtakomuníkacíjnihmerežah AT novogrudskarina ontologíčnamodelʹdlâorganízacííprocesuobroblennâdanihvínformacíjnihtakomuníkacíjnihmerežah |