An Ontology-based Security Framework for Decision-making in

Industrial Systems

Bruno A. Mozzaquatro

, Raquel Melo

, Carlos Agostinho

and Ricardo Jardim-Goncalves

Universidade Nova de Lisboa (UNL), DEE/FCT, 2829-516 Caparica, Portugal

Centre of Technology and Systems, UNINOVA, 2829-516 Caparica, Portugal

Keywords:

Ontology, Internet of Things, Information Security, Decision-Making System, Adaptive Security.

Abstract:

Embedded devices based on emerging technologies of the Internet of Things (IoT) are used to provide re-

sources, business models and opportunities to build potential industrial systems improving manufacturing

systems with efﬁcient operations. In this context, IoT networks are dynamic environments and changes are

also being increasingly frequent, modifying the environment execution. Nevertheless, severe threats will in-

crease the complexity and difﬁculty to protect existing vulnerabilities in smart devices of IoT network. In

this context, this work proposes an architecture of the ontology-based security framework to decision-making

using adaptive security model to improve secure information for the industrial systems. IoTSec ontology con-

tributes to feed the system using queries of contextual information collected in the environment. The main

contribution of this approach is validated as an integration with C2NET project to ensure security properties

in some critical scenarios.

1 INTRODUCTION

Smart devices integrated with different technologies

allow several industrial applications with sensing,

identiﬁcation, localization, networking and process-

ing capabilities. The information technology (IT)

standards have beneﬁciated the industrial manufactur-

ing by the evolution of industrial systems (Bi et al.,

2014). The adoption of the Internet creates new busi-

ness opportunities as well as exploiting collaborative

work based on IT infrastructure in system environ-

ments. These aspects have potential to develop indus-

trial systems like environmental monitoring, health-

care service, inventory and production management,

food supply chain, transportation, workplace and

home support, security and surveillance (Xu et al.,

2014).

Heterogeneous environments with smart devices

interconnected with the Internet increases the secu-

rity threats. The main problem of IoT security is

high interaction between humans, machines and IoT

technologies with constraints in terms of connectivity,

computational power and energy (Sicari et al., 2014).

In contrast, severals security models, trust manage-

ment, identity management and security mechanisms

are used to ensure the privacy and security keeping se-

curity goals, such as: availability, conﬁdentiality, in-

tegrity, authentication, non-repudiation, and authen-

ticity (Roman et al., 2013) (Yan et al., 2014) (Gran-

jal et al., 2014). Furthermore, IoT network is a dy-

namically changing environment and security issues

require making-decision systems to change security

mechanisms at run time (Evesti and Ovaska, 2013).

Therefore, the adaptive security model is necessary to

learn and adapt for adjusting on-demand security at-

tributes and antecipates new threats in an information

system (Habib and Leister, 2013).

Information security is an important requirement

to fully adoption of IoT applications and must be con-

sidered by information system designers and by ad-

ministrators of organizations that depends on the cor-

rect management of information security and conﬁ-

dentiality (Yan et al., 2014). However, IoT is still in a

conceptual phase, but the ﬁeld is very dynamic and se-

curity challenges are less structured, somewhat orga-

nized causing confusion amongst concepts and terms

to software developers. Ontology characterizes an in-

terest domain with classes and relationships among

them and implements a data model to share a com-

mon base knowledge in the particular domain (Moz-

zaquatro et al., 2015).

Model-Driven Development (MDD) has relevant

aspects that contributes to develop adaptive systems

considering adaptive model to monitor contextual in-

formation and take suitable actions at runtime (Soylu

Mozzaquatro, B., Melo, R., Agostinho, C. and Jardim-Goncalves, R.

An Ontology-based Security Framework for Decision-making in Industrial Systems.

DOI: 10.5220/0005853107790788

In Proceedings of the 4th International Conference on Model-Driven Engineering and Software Development (MODELSWARD 2016), pages 779-788

ISBN: 978-989-758-168-7

779

Existing Security Ontologies

(State of the art)

Security overview Specific Domain

Herzog et al.

Fenz, S. et al.

Kim, A. et al.

Denker et al.

Undercoffer, J. et al.

Intrusion Detection

Gyrard, J. et al.

M2M Architecture

Garca-Crespo, A. et al.

Access Control Policies

Mozzaquatro, B. et al.

IoT Security

Figure 1: Existing security ontologies.

and De Causmaecker, 2009). In this context, the re-

lation of MDD with Ontology Driven Development

(ODD) employ the use of formal model to be em-

ployed at runtime. Hence, the application of on-

tologies for information security could improves real

time detection of vulnerabilities, prediction and as-

sessment of security risk management and intrusion

detection (Undercoffer et al., 2003) (Xu et al., 2009)

(Frye et al., 2012). Reference ontology in the secu-

rity community has been identiﬁed as an important

challenge (Mouratidis, 2006) and also for the Internet

of Things. The previous work (Mozzaquatro et al.,

2015) was proposed a reference ontology for secu-

rity in the IoT with harmonization of existing security

ontologies based on ontology development methodol-

ogy. In addition, a security framework is fundamen-

tal to make secure the IoT environment allowing do

queries and inferences to the security issues.

In this work we propose an architecture for an

ontology-based adaptive security framework to iden-

tify common security issues using a knowledge base

and demonstrate the contribution of application of the

framework based on two security approaches: design

and run time. This framework explores the knowledge

base of IoTSec ontology (Mozzaquatro et al., 2015) to

realize queries according to the contextual informa-

tion collected of the smart environment in industrial

scenario.

The rest of paper is organized as follow: Section

2 presents the background of security ontologies and

adaptive security model. The related works are pre-

sented in Section 2.3. Section 4 describes the archi-

tecture proposed for ontology-based security frame-

work. Section 5 describes a case study of the contri-

bution of the approach. Finally, Section 6 presents the

conclusion about this work.

2 BACKGROUND

This section describes main subjects involved in this

work that contributes to improve the security aspects

in the context of Internet of Things. In the following

we discuss existing security ontologies (i.e. IoTSec

ontology) and the adaptive security model to adapt

mechanisms in a suitable solution. In addition, related

works are presented to demonstrate the originality of

the paper.

2.1 Security Ontologies

Security issues are important for all contexts with per-

sonal data exchanges and sensitive information, but

for Internet of Things has important characteristics

of a big concern with high iteration between humans,

machines and IoT technologies. It is justiﬁed by het-

erogeneity of different smart devices connected with

the Internet. In this context, ensuring security and pri-

vacy of applications and services is critical to improv-

ing trust and use of the Internet.

Therefore, these problems have potential because

there are several situations of misunderstood concepts

around information security and Internet of Things.

For that, ontology is a potential tool largely utilized

for structuring an area of interest.

According to the state of the art, several existing

security ontologies have been proposed in the litera-

ture, but only a few are available (Figure 1): secu-

rity overview ontology (Herzog et al., 2007) (Fenz

and Ekelhart, 2009) (Kim et al., 2005) (Denker et al.,

2003) and security ontology applied to speciﬁc do-

main (Undercoffer et al., 2003) (Gyrard et al., 2014)

(Garc

ıa-Crespo et al., 2011) (Mozzaquatro et al.,

2015).

MDE4SI 2016 - Special Session on Model-Driven Enterprise Services and Applications for a Sustainable Interoperability: New Paradigms

for Development in the Future Enterprise - 2nd Edition

780

Asset%

Threat%

Vulnerability%

SecurityProperty%

threatens%

SecurityMechanism%

protects%

hasVulnerability%

isSecurityMechanismOf%

aﬀects%

Feature%

hasFeature%

requires%

SeverityScale%

hasSeverity%

TypeOfDefense%

hasType%

OSIModel%

occursIn%

Figure 2: Reference ontology for security in IoT (Mozzaquatro et al., 2015).

Some ontologies address only one part of secu-

rity domain (e.g. computer attacks) and others ex-

plore overview of information security. The previous

work (Mozzaquatro et al., 2015) proposed a reference

ontology for security in the IoT (IoTSec) with harmo-

nization of ontologies based on ontology development

methodology.

2.1.1 IoTSec Ontology

In this section, we describe the IoTSec ontology,

which is a reference ontology for IoT security pro-

posed in (Mozzaquatro et al., 2015). IoTSec on-

tology was proposed to explore aspects of relation-

ships among basic components of the risk analysis

of ISO/IEC 13335-1:2004 and National Institute of

Standards and Technology (NIST) Special Publica-

tion 800-12 (Stoneburner et al., 2002) such as: As-

sets, Threats, SecurityMechanism, Vulnerability and

Risk. Figure 2 presents an arrangement of top-level

classes to modeling information security based in

works (Herzog et al., 2007) (Fenz and Ekelhart, 2009)

(Kim et al., 2005) (Denker et al., 2003) (Gyrard et al.,

2014).

IoTSec ontology was designed based on informa-

tion security issues that can be represents using a

structured knowledge. Basically, ontology explores

relationships among classic components of risk anal-

ysis to provide an overview of the domain of security

in Internet of Things. IoTSec ontology was developed

using the OWL (Web Ontology Language) ontology

language.

These components allow to identify relations be-

tween relevant situations in an IoT network with risk

analysis of potential threats. For example, the vul-

nerability class describes potential weakness of M2M

technologies associated with Asset class (hasVulner-

ability property). In this ontology many technolo-

gies are considered assets such as: Wi-Fi, Web, GSM

(2G), UTMS (3G), LTE (4G), Ethernet, Bluetooth,

Sensor, etc. Assets require security properties to be

considered secure such as availability, conﬁdential-

ity, integrity, etc. Vulnerabilities are ﬂaws in software

or hardware and when they are discovered, vendors

publish a patch to ﬁx it. For instance, vulnerabil-

ity notes database (VND)

is one example that pro-

vides information about software vulnerabilities in-

cluding summaries, technical details, remediation in-

formation, and lists of affected vendors.

Meanwhile, security mechanisms are used to

avoid that threats exploit vulnerabilities found. These

mechanisms are categorized according to type of de-

fense to protect the assets. A security mechanism is

composed of several types of defense i.e., detective,

preventive, corrective, recovery, response, etc.

Threat class describes information about attacks

and others ways to exploit the applications’ weakness

and, sometimes, they explore one or more vulnerabili-

ties. For instance, Wormhole attack replays messages

from a system with the vulnerability of unprotected

communication channel of a sensor network of an or-

ganization. This threat occurs in network layer of OSI

Model (occursIn property). In this situation, Secu-

rityMechanism class contains tools to protect using

cryptography algorithms, but need to consider their

strengths and weaknesses such as energy consump-

tion, ﬂexibility, high cost, etc.

Organizations may prevent exploitation of your

vulnerabilities using security tools or algorithms to

protect (mitigates property) the systems’ weakness.

Mitigates property represents the relationship be-

tween SecurityMechanism and Vulnerability class.

Vulnerabilities are qualiﬁed in terms of your severity

level (SeverityScale class) to an organization. Some-

times, organizations need to monitor the vulnerability

with severity scale high, when systems have behavior

http://nvd.nist.gov/

An Ontology-based Security Framework for Decision-making in Industrial Systems

781

unpredictable and they can have become exposed to

new threats. Each threat affects one or more security

properties and the security mechanisms could satisfy

these security properties.

2.2 Adaptive Security

Adaptive security is an approach to adjust attributes

based on the behavior at runtime to respond for new

and unusual threats in critical services (Abie, 2009)

(Habib and Leister, 2013). This approach is found

in the literature with concepts of self-adaptive soft-

ware (Laddaga and Robertson, 2004) and autonomic

computing (Dobson et al., 2006). It is a solution

that learns and adapts to the changing environment

at runtime in face of changing threats, and anticipates

threats before they are manifested.

This approach is a continuous process to learn,

adapt, prevent, identity and respond for unusual and

malicious behavior in run time. For that, the adaptive

security model proposed by (Shnitko, 2003) is com-

posed by four components as depicted in Figure 3:

monitor, analyzer, adapter, and adaptive knowledge

database. The monitor collects attributes, analyzer

determines the adaptation requirements, and adapter

decides the adaptation plan for execution.

Figure 3: Adaptive and evolving security model (Shnitko,

2003).

The continuous cycle of security monitoring is

needed for the use of suitable mechanisms depend-

ing of the information about the context and status of

IoT devices. It is appropriated for IoT scenario be-

cause high interactions among heterogeneous devices

and environment with critical risks to our lives. Moni-

toring information context allows to choose a suitable

security tool for ensure one or more security prop-

erties. Sometimes, whether it changes the status of

secure behavior of a given situation, the system need

to change your rules to adapt and ensuring security

properties.

According to (Habib and Leister, 2013) there is a

little work on adaptive security mechanisms to secure

IoT. Each work proposed explores platforms and spe-

ciﬁc aspects to improve IoT security as well as secu-

rity policies, encryption, secure communication, and

intrusion detection. Basically, there is a need for IoT

security to adapt and adjust attributes when there is

a change in the context. Nevertheless, the reliabil-

ity and performance of adaptive security approaches

is directly related with security mechanisms used to

identify the threats in the system.

Within the scope of Model-Driven Development

(MDD), adaptive security could be addressed to pro-

vides customization as a service in a runtime archi-

tecture. MDD is an approach composed by several

theories and methodological frameworks for indus-

trialized software development using models inside

of software development cycle (Picek and Strahonja,

2007). There models and your transformations are de-

scribed based on standard speciﬁcation languages and

generated automatically or semi-automatically from

others abstract models. One relevant aspect of MDD

to the adaptive security is the automation, which non-

code artifacts are produced totally or partially from

models (Picek and Strahonja, 2007) such as: docu-

mentation, test artifacts, build and deployment scripts

and other models.

Adaptive actions are mediated by automated pro-

cess through systems to maintain a formal model (i.e.

context-aware) of the settings and relationships be-

tween them(Soylu and De Causmaecker, 2009). Be-

sides it, model-driven approach can also be consid-

ered as an ontology driven approach, but the integra-

tion of these two approaches migh result beneﬁts of

inference support of ontological approaches and the

expertise of model driven approach. Therefore adap-

tive actions are beneﬁciated with transformation of

OWL/RDF knolwedge base into domain-centric data

models (Kalyanpur et al., 2004).

2.3 Related Works

Ontologies has been explored in several aspects to im-

prove information security, identifying vulnerabilities

of systems, assessment the threat against targets us-

ing differents approaches, such as: intrusion detec-

tion (Undercoffer et al., 2003), correlation of context-

aware alert analysis (Xu et al., 2009), identiﬁcation of

complex network attacks (Frye et al., 2012).

The work (Evesti and Ovaska, 2013) proposes an

architectural approach for security adaptation in smart

spaces utilized to analyzing and planning access con-

trol decision at runtime and design-time. The au-

thors combines an adaptation loop of adaptive secu-

MDE4SI 2016 - Special Session on Model-Driven Enterprise Services and Applications for a Sustainable Interoperability: New Paradigms

for Development in the Future Enterprise - 2nd Edition

782

rity model, Information Security Measuring Ontol-

ogy (ISMO) to offer input knowledge for the adaption

loop and a smart space security-control model to en-

force dynamic access control policies. However, the

work only illustrates the adaptive security approach

from the authentication and authorization of users of

smart spaces.

EDAS (Aman and Snekkenes, 2014) was pro-

posed as an event driven adaptive security model to

IoT to protect devices against threat faced at runtime.

The authors use an Open Source Security Informa-

tion Management (OSSIM) to ﬁlter and normalize

events collected from things. They explore an Adap-

tation Ontology to leverages risks information from

the event correlation and adapt security settings in

terms of usability, QoS, and security reliability. How-

ever, the authors do not consider potential vulnerabili-

ties that could prevent eventual threats in the environ-

ment. In this case, the approach need an occurrence

to verify the suitable action to mitigate it.

In this context, this work explores a reference on-

tology for security in the Internet of Things. These

knowledge based is composed by the basic compo-

nents of risk management to ensure the secure en-

vironment with IoT devices. The relation between

threat, vulnerability, asset, security mechanism and

security property enable to make decision for use of

potential solutions or identify weakness of products

or softwares in industrial systems, for example. These

ontology-based decision-making approach has poten-

tial to enrich security mechanisms of IoT devices net-

work using adaptive security model through to re-

spond for unusual behavior proposing other security

tool or algorithms to protect assets.

3 C2NET PLATFORM

The C2NET platform is cloud-enabled tools for sup-

porting collaborative demand to cover the supply net-

working optimization of manufacturing and logistic

assets. The main problem of traditional supply chains

has centralized decision-making approaches, which

make difﬁcult for companies to react to current highly

dynamic markets. According it, C2NET platform is

proposed to contributes in several aspects of indus-

trial manufacturing exploring data collection of IoT

devices in the companies’ shop ﬂoor.

However, these devices are vulnerable for several

threats and it needs to be addressed using security

mechanisms. Moreover, some of these devices use

different IoT technologies and C2NET platform ex-

plores the interoperability based on semantic web’

technologies.

The C2NET architecture ensures interoperabil-

ity by deﬁning two components: C2NET Agent

and C2NET Data Collection Client. Moreover, the

C2NET platform has a module to collecting data from

different sources and provide support for others mod-

ules of C2NET system.

3.1 C2NET Data Collection Client

The C2NET Data Collection Client (DCC) is a com-

ponent of C2NET platform that provides the collect-

ing and sending all the required data from the legacy

systems of the company (e.g. its planning, logistics

and operations) and data arriving from IoT devices

in the shop ﬂoor (e.g. machine availability, perfor-

mance, etc).

This component must be able to connect the dif-

ferent data sources (both legacy systems and IoT de-

vices). Then, it will store the data gathered, and sub-

mit it to C2NET DCF as events when needed (both in

a periodical basis or under demand). Consequently,

the C2NET DCC will adopt an ESB pattern.

3.2 C2NET Agent

The C2NET Agent is a component of C2NET plat-

form that receives the information generated in the

platform (e.g. new collaborative production plans)

for transferring it to the systems of the companies in-

volved in the value chain. The C2NET platform will

use the Agent API to communicate with the C2NET

Agent, which exposes the legacy systems as busi-

ness services. The C2NET Agent could receive the

C2NET message and call the proprietary API of each

legacy system to perform data update as needed.

3.3 C2NET Data Collection Framework

The C2NET Data Collection Framework (DCF) is a

domain module that offers functionality for collecting

data from different heterogeneous sources and pro-

viding necessary information for other modules of

the C2NET system. This module enables uniform

accessibility of structured information for data con-

sumers. It also resolves challenges concerning inte-

gration and interoperability across data producers and

consumers caused by differences in industrial pro-

cesses, data models, methods, technologies and de-

vices. It takes into consideration the homogenous in-

tegration of legacy systems and IoT devices.

An Ontology-based Security Framework for Decision-making in Industrial Systems

783

Figure 4: An architecture of ontology-based security framework proposed with the C2NET platform.

4 ONTOLOGY-BASED SECURITY

FRAMEWORK

We develop our ontology-based security framework

by following the adaptive security model presented in

the Section 2.2. This type of security model is suitable

for dynamic environment of Internet of Things, which

monitor the behavior of the environment, learns and

adapts for unusual occurences. In this context, the

support of an adaptive knowledge base enables to an-

tecipates threats before they are manifested in the IoT

network.

Several industries are using of IoT devices to

different applications to provides new opportunities

based on sensing, ubiquitous identiﬁcation, and com-

munication capabilities (Xu et al., 2014). IoT devices

transmit sensitive information of companies and they

are vulnerable to internal and external attacks and ap-

propriates actions need to be adopted to protect them.

The ontology-based security framework proposed

in this work explores the adaptive security model to

making-decision based on knowledge base for in-

formation security issues. For that, this security

framework is integrated with the platform of C2NET

project

to enrich security of IoT devices in industrial

systems. In this context, the architecture of ontology-

based security framework is depicted in the Figure 4.

The architecture aims to improve security issues of

industrial manufacturing integrated with the C2NET

http://c2net-project.eu

platform.

The C2NET platform uses IoT devices to data col-

lect of the industrial environment using a company

middleware with the C2NET Data Collection. Secu-

rity mechanisms (i.e. based on rules, security pro-

tocols) are applied in data communication to protect

sensitive information between IoT devices and mid-

dleware. Nevertheless, several vulnerabilities of de-

vices and software appear everyday, which they be-

coming the assets vulnerable to attacks. Hence, con-

tinuous assessment and suitable adaptations need to

be enforced to ensure the security properties such as

availability, conﬁdentiality, and integrity.

The security framework is proposed with two ap-

proaches to improve security issues of C2NET plat-

form: design and run time. Design approach of the

security framework explores the previous knowledge

to adopt new technologies or products considering se-

curity issues. It has impact in the companies, be-

cause the responsible of purchases have not expertise

in information security and it becomes the purchase

of product without security analysis.

On the other hand, run time approach monitors

IoT devices based on security metrics and attributes

to identify malicious behaviors in the smart environ-

ment. Consequently, conﬁgurations and/or rules need

to be adapted according the knowledge base, when

alerts are triggered by security tools. For that, IoT on-

tology contributes to identify relations between threat,

asset, vulnerability, security mechanism and security

property. Nevertheless, the adapter infers in new

MDE4SI 2016 - Special Session on Model-Driven Enterprise Services and Applications for a Sustainable Interoperability: New Paradigms

for Development in the Future Enterprise - 2nd Edition

784

information on knowledge base to deploy new ap-

proaches for speciﬁcs situations or malicious behav-

iors.

Considering the approaches of security frame-

work, this architecture is divided in two main compo-

nents that are integrated with components of C2NET

platform which are DCC and DCF. Design approach

uses an interface to realize queries in C2NET DCF

for data analysis based on the IoTSec ontology. This

interface is used also to update the knowledge base

when new information was published.

By other hand, the runtime approach has a Real

Time Reasoning module to identify anomaly be-

haviour based on security mechanisms used in the

environment. In this context, security attributes are

monitored to detect malicious activities and, then, an

action is trigerred to adjust settings using the adaptive

security API of security framework.

4.1 Monitor Module

The Monitor module is responsible to data collection

of information between IoT devices and enforce some

suitable rules to the IoT newtork. These devices are

used to information gathering of security attributes of

environment. The monitor only considers device’s in-

formation collected by devices to identify potential

vulnerabilities that could be explored by the threats.

Hence, raw data of the environment (e.g. shop ﬂoor)

also is collected, but this information only is ﬁltered

by the platform to verify the proper operation based

in the semantic of data.

This module uses a temporary set of security rules

to apply of ensure a secure data communication be-

tween IoT devices. This approach is used in run time

to make decision without delay. In this case, this set

is feeded with previous knowledge of the ontology ac-

cording to device’s information of security attributes

collected. For that, the Real Time Analysis module

works together with monitor module to analyze pa-

rameters upward of threshold deﬁned. For instance,

any unusual behavior identiﬁed by security tools like

intrusion detection system or ﬁrewall are fowarded to

DCF component of C2NET platform to realize the

most deep analysis of ontology-based security frame-

work.

4.2 Analyzer Module

The Analyzer module of the security framework is

responsible for consulting activities mapped in the

knowledge base, but in case of new occurences (e.g.

zero-day threats) reported by security tools, resulting

in unusual behavior for the security framework. So,

it needs to be adapted to avoid critical harms to the

organizations.

DCF component of the C2NET platform manages

the virtual instances of IoT devices to control your be-

havior in physical world. The Resource Virtualization

is resposible to minimize the distance between phys-

ical devices and their virtualized devices in the IoT

network. This information is important to analyzer

module make decision for adaptation for potential se-

curity solutions between IoT devices.

4.3 Adapter Module

The Adapter module uses different adaptation meth-

ods to reconﬁguration mechanisms employed by the

adaptive security contained in the security frame-

work. More information of this adaptation methods

are found in (Elkhodary and Whittle, 2007).

5 ONTOLOGY-BASED SECURITY

FRAMEWORK APPLICATION

In this section, we describe two validation scenarios

of metalworking industry to apply the ontology-based

security framework to improve the security issues be-

tween IoT devices and C2NET platform.

The IoTSec ontology was designed to allow

decision-making following main classes of security:

assets (K and X), vulnerabilities (Y), threats (Z), se-

curity mechanisms (G), and security properties. For

instance, a sensor K (K requires some security prop-

erties J) based on technology X has vulnerabilities Y

that can threatens by attack Z, but if the company use

a security mechanism G, he could neutralize poten-

tial threats. Also, organizations need to deﬁne restric-

tions that can be represented by security policies and

it means the security properties addressed for security

framework.

To demontrate the application of security frame-

work and relationship between IoTSec ontology and

adaptive security model, Figure 5 presents a step-by-

step application of framework.

• 1

◦

step: The security attributes are collected by

the environment to identify potential threats or

security mechanisms to protect data communica-

tion, for example.

• 2

◦

step: A temporary set of security rules is used

to apply mechanism to protect devices and data

transmissions. It is used for run time approach,

which requires a fast decision-making based in

previous knowledge.

An Ontology-based Security Framework for Decision-making in Industrial Systems

785

Figure 5: Step-by-step application of security framework.

• 3

◦

step: In case of unusual behavior or a deep

analysis is required to ﬁnd other solutions. Thus,

knowledge base is checked to identify suitable se-

curity mechanism that is related with a vulnerabil-

ity.

• 4

◦

step: SPARQL queries are used to collect infor-

mation about an unusual behavior or a new vulner-

ability, for example. In this case, IoTSec ontology

gives support for decision-making using security

attributes collected by IoT devices.

• 5

◦

step: Several situations are not mapped in

knowledge base and it needs to be adapted using

differents types of adaptation methods.

• 6

◦

step: After a deep analysis the adaptation

method update the knowledge base to future oc-

curences.

We describe an overview of each scenario to un-

derstand the applicability of security framework and

what information are important to monitoring for se-

curity level assessment. According to the restric-

tions and contextual information, security framework

checks the knolwedge base using SPARQL Protocol

and RDF Query Language (SPARQL)

queries and

relates with others queries to identify suitable security

mechanisms or potential threats, for example. Hence,

in this work we considered that scenarios are vulner-

able only for digital threats, such as disclosure infor-

mation, replay attack, spooﬁng and others attacks to

smart devices.

5.1 Collaborative Purchase

Collaborative purchase is a validation scenario that

the C2NET platform allows to subscribe to common

purchases on the same suppliers for the companies’

partners. Collaborative purchase is need to found

other companies that require same products of a spec-

iﬁed supplier. Information of products that are miss-

ing are collected by using the IoT devices about the

speciﬁcations and features of the raw material re-

quired (such as material type, material code, dimen-

sions or weight), and also provides an expiration date

http://www.w3.org/TR/rdf-sparql-query/

for the same. This expiration date is related with

the expected deadline for the purchase that is directly

connected to the date that the company manager needs

to have the product in his company. Moreover, smart

sensors are used to verify the availability of the raw

material and offer a collaborative purchase for others

companies with the same supplier.

Metalworking industry imposes severe restric-

tions in the production process and one of them is

the availability of raw material required to main-

tain it working. In this context, security property of

availability of the sensors to gathering information

about the products that are missing is fundamental

to the process. The query following selects the se-

curity mechanisms that satisﬁes the security property

“Availability”. The variable ?secmech represents se-

curity mechanisms contained in the knowledge base

and they are related with security property (variable

?secprop).

SELECT DISTINCT ?secmech ?secprop

WHERE

?secmech iotsec:satisfies ?secprop .

?secprop rdfs:label "Availability"@en

Moreover, authentication is another security prop-

erty responsible to identify who has access to the in-

formation and avoid unauthorized user using these

data. Sometimes, malicious employees or users using

the Internet could explore vulnerabilities to realized

attacks causing information disclosure or others crit-

ical consequences. For example, if a sensor has vul-

nerability that allows attackers get access of internal

network it needs to be ﬁxed. Hence, the ontology-

based security framework has important role to check

the relation between vulnerabilities and threats based

on the knowledge base (IoTSec ontology). The query

following uses the variable ?threat that threatens

vulnerabilities selected by variable ?vuln.

SELECT DISTINCT ?threat ?vuln

WHERE

?threat rdfs:label ?label .

?threat iotsec:isVulnerabilityOf ?vuln

MDE4SI 2016 - Special Session on Model-Driven Enterprise Services and Applications for a Sustainable Interoperability: New Paradigms

for Development in the Future Enterprise - 2nd Edition

786

Security attributes of IoT devices the framework

could suggest defense tools to block malicious be-

havior or attacks using prior knowledge. In another

way, new activities are adapted with adjusting internal

working parameters such as encryption schemes, al-

gorithms of intrusion detection systems, different au-

thentication and authorization mechanisms.

5.2 Non-conformity Scheduling

Non-conformity scheduling is a validation scenario

composed of several IoT devices/sensors to feed the

C2NET platform with real time detection of non-

conformity products during the production process.

It reduces the quantity of waste and non-conformity

products that may arise during production process.

This scenario is the most critical in case of secu-

rity issues because the C2NET platform collects in-

formation about the shop-ﬂoor production. Then, any

violation in data communication between IoT devices

and C2NET platform could compromise the produc-

tion or to result critical problems to the company.

In this scenario the main security properties that

need to be ensured are availability, integrity, conﬁden-

tiality, and authentication. The availability property

is recommended to maintain all information accessi-

ble to the C2NET platform as well as the continuity

of production process. The integrity and conﬁden-

tiality properties are two important aspects to avoid

access to the information transmitted. For that, se-

curity framework contributes to choose suitable se-

curity tools with the relation between vulnerabilities,

threats and security mechanisms. Basically, the query

following shows how this information is obtained of

IoTSec ontology, considering the variables ?threat,

?secmec (security mechanism) and ?secprop (secu-

rity property).

SELECT DISTINCT ?threat ?secmec ?secprop

WHERE

?threat rdfs:label ?label .

?threat iotsec:hasSecurityMechan ?secmec .

?secmec iotsec:satisfies ?secprop

In case of conﬁdentiality, if an attacker has

got access to the information transmitted, he can

not understand this information because it must be

encrypted. Moreover, the authentication property is

related only with access control of the information by

the C2NET platform.

6 CONCLUSION

We presented an architecture of ontology-based se-

curity framework to decision-making systems using

adaptive security model to improve security issues in

industrial scenario. In this context, the IoTSec ontol-

ogy is responsible to shows a representation of struc-

tured knolwedge using semantic web technologies in

the context of information security.

We have demonstrate how our proposed architec-

ture ensure information security and there is poten-

tial to maintain security requirements need for risk

management using main components of risk analysis.

Merge between MDD and ODD approaches has po-

tential to improve the reasoning and adaptive actions

based on the knowledge base using contextual infor-

mation.

After the integration of security framework with

C2NET platform, we intend to continue checking new

adaptive security models to replace or complement

these already proposed; checking new knolwedge

bases related of Internet of Things and M2M commu-

nication to specialize according to speciﬁc scenarios.

ACKNOWLEDGEMENTS

The research leading to this work has received fund-

ing from CAPES Proc. N

: BEX 0966/15-0 and

European Commission’s Horizon 2020 Programme

(H2020/2014-2020) under grant agreement: C2NET

: 636909.

REFERENCES

Abie, H. (2009). Adaptive security and trust management

for autonomic message-oriented middleware. Mobile

Adhoc and Sensor Systems, 2009. MASS ’09. IEEE 6th

International Conference on, pages 810–817.

Aman, W. and Snekkenes, E. (2014). Event driven adaptive

security in internet of things. pages 7–15.

Bi, Z., Xu, L. D., and Wang, C. (2014). Internet of things for

enterprise systems of modern manufacturing. IEEE

Transactions on Industrial Informatics, 10(2):1537–

1546.

Denker, G., Kagal, L., Finin, T., and Paolucci, M. (2003).

Security for daml web services : Annotation and

matchmaking. pages 335–350.

Dobson, S., Zambonelli, F., Denazis, S., Fern

andez, A.,

Ga”ıti, D., Gelenbe, E., Massacci, F., Nixon, P., Saf-

fre, F., and Schmidt, N. (2006). A survey of autonomic

communications. ACM Transactions on Autonomous

and Adaptive Systems, 1(2):223–259.

An Ontology-based Security Framework for Decision-making in Industrial Systems

787

Elkhodary, A. and Whittle, J. (2007). A survey of ap-

proaches to adaptive application security. In Proceed-

ings of the 2007 International Workshop on Software

Engineering for Adaptive and Self-Managing Systems,

page 16. IEEE Computer Society.

Evesti, A. and Ovaska, E. (2013). Comparison of adaptive

information security approaches. ISRN Artiﬁcial In-

telligence, 2013.

Fenz, S. and Ekelhart, A. (2009). Formalizing informa-

tion security knowledge. Proceedings of the 4th Inter-

national Symposium on Information, Computer, and

Communications Security - ASIACCS ’09, page 183.

Frye, L., Cheng, L., and Heﬂin, J. (2012). An ontology-

based system to identify complex network attacks.

IEEE International Conference on Communications,

pages 6683–6688.

Garc

ıa-Crespo,

A., G

omez-Berb

ıs, J. M., Colomo-Palacios,

R., and Alor-Hern

andez, G. (2011). Securontology:

A semantic web access control framework. Computer

Standards & Interfaces, 33(1):42–49.

Granjal, J., Monteiro, E., and Silva, J. S. (2014). Security

in the integration of low-power wireless sensor net-

works with the internet: A survey. Ad Hoc Networks,

24:264–287.

Gyrard, A., Bonnet, C., and Boudaoud, K. (2014). An

ontology-based approach for helping to secure the etsi

machine-to-machine architecture. IEEE International

Conference on Internet of Things 2014 (iThings).

Habib, K. and Leister, W. (2013). Adaptive security for

the internet of things reference model. Norsk infor-

masjonssikkerhetskonferanse (NISK), pages 13–25.

Herzog, A., Shahmehri, N., and Duma, C. (2007). An on-

tology of information security. Journal of Information

Security, 1(4):1–23.

Kalyanpur, A., Pastor, D. J., Battle, S., and Padget, J. A.

(2004). Automatic mapping of owl ontologies into

java. In SEKE, volume 4, pages 98–103. Citeseer.

Kim, A., Luo, J., and Kang, M. (2005). Security ontol-

ogy for annotating resources. In On the Move to

Meaningful Internet Systems 2005: CoopIS, DOA,

and ODBASE, pages 1483–1499.

Laddaga, R. and Robertson, P. (2004). Self adaptive soft-

ware : A position paper. SELF-STAR: International

Workshop on Self-* Properties in Complex Informa-

tion Systems, 19:31.

Mouratidis, H. (2006). Integrating Security and Software

Engineering: Advances and Future Visions: Advances

and Future Visions. IGI Global.

Mozzaquatro, B. A., Jardim-goncalves, R., and Agostinho,

C. (2015). Towards a reference ontology for security

in the internet of things. In IEEE International Work-

shop on Measurement and Networking, pages 1–6.

Picek, R. and Strahonja, V. (2007). Model driven

development-future or failure of software develop-

ment. In IIS, volume 7, pages 407–413.

Roman, R., Zhou, J., and Lopez, J. (2013). On the features

and challenges of security and privacy in distributed

internet of things. Computer Networks, 57(10):2266–

2279.

Shnitko, A. (2003). Adaptive security in complex informa-

tion systems. In Science and Technology, 2003. Pro-

ceedings KORUS 2003. The 7th Korea-Russia Inter-

national Symposium on, pages 206–210.

Sicari, S., Rizzardi, a., Grieco, L., and Coen-Porisini, a.

(2014). Security, privacy and trust in internet of

things: The road ahead. Computer Networks, 76:146–

164.

Soylu, A. and De Causmaecker, P. (2009). Merging model

driven and ontology driven system development ap-

proaches pervasive computing perspective. In Com-

puter and Information Sciences, 2009. ISCIS 2009.

24th International Symposium on, pages 730–735.

IEEE.

Stoneburner, G., Goguen A. Y., and Feringa, A. (2002). Sp

800-30. risk management guide for information tech-

nology systems.

Undercoffer, J., Joshi, A., and Pinkston, J. (2003). Model-

ing computer attacks : An ontology for intrusion de-

tection. pages 113–135.

Xu, H., Xiao, D., and Wu, Z. (2009). Application of se-

curity ontology to context-aware alert analysis. 2009

Eighth IEEE/ACIS International Conference on Com-

puter and Information Science, pages 171–176.

Xu, L. D., He, W., and Li, S. (2014). Internet of things in

industries: A survey. IEEE Transactions on Industrial

Informatics, 10(4):2233–2243.

Yan, Z., Zhang, P., and Vasilakos, A. V. (2014). A survey

on trust management for internet of things. Journal of

Network and Computer Applications, 42(2):120–134.

MDE4SI 2016 - Special Session on Model-Driven Enterprise Services and Applications for a Sustainable Interoperability: New Paradigms

for Development in the Future Enterprise - 2nd Edition

788