Page 163 - Proceedings of the 2017 ITU Kaleidoscope

DATA CENTRIC TRUST EVALUATION AND PREDICTION
FRAMEWORK FOR IOT

2
1
1
1
Upul Jayasinghe , Abayomi Otebolaku , Tai-Won Um , Gyu Myoung Lee
1 Department of Computer Science, Liverpool John Moores University, Liverpool, L3 3AF, UK.
2 Department of Information and Communication Engineering, Chosun University, Gwangju, Korea.
u.u.jayasinghe@2015.ljmu.ac.uk, a.m.otebolaku@ljmu.ac.uk, twum@chosun.ac.kr, g.m.lee@ljmu.ac.uk
ABSTRACT reliable, up to date and location sensitive information about
weather, traffic, safety warnings and transport information
Application of trust principals in internet of things (IoT) has from a smart city application are more important than the
allowed to provide more trustworthy services among the facts about entities who are actually generating them. The
corresponding stakeholders. The most common method of other common misinterpretation is that the assumption of
assessing trust in IoT applications is to estimate trust level having entity trust would guarantee data trust which is in fact
of the end entities (entity-centric) relative to the trustor. In indubitably different in various aspects like validity of data,
these systems, trust level of the data is assumed to be the timeliness and other properties unique to data which are
same as the trust level of the data source. However, most of often ignored in calculating trust for end entities. Further,
the IoT based systems are data centric and operate in information is the governing factor for any IoT systems and
dynamic environments, which need immediate actions is generated from the data by combining it (data) with the
without waiting for a trust report from end entities. We context. Hence, if there is a data quality (DQ) problem, it
address this challenge by extending our previous proposals would eventually lead to information quality (IQ) problem
on trust establishment for entities based on their reputation, [22]. In other words, once the right data item is delivered to
experience and knowledge, to trust estimation of data items a desired entity at the precise time in a clear, useable and
[1-3]. First, we present a hybrid trust framework for meaningful manner, IQ is guaranteed.
evaluating both data trust and entity trust, which will be
enhanced as a standardization for future data driven society. Therefore, it is important to address the challenge of
The modules including data trust metric extraction, data establishing a data centric trust while preserving the
trust aggregation, evaluation and prediction are elaborated traditional form of trust computation. To this end, firstly we
inside the proposed framework. Finally, a possible design define a set of dynamic factors, which essentially describe
model is described to implement the proposed ideas. the DQ attributes and also metrics which define the
knowledge, experience and reputation as in our previous
Keywords— Data Trust, Knowledge, Reputation, work [2] and [1] to get the best of traditional means of trust
Experience, Collaborative Filtering, Ensemble Learning. computation. Then, we combine these attributes built on
REK (Reputation, Experience and Knowledge) model
1. INTRODUCTION described in [4] and [7]. After that, a technique which
assesses the data centric trust for every user who is new to
With the exponential growth of applications of internet of the system and who needs to access the data streams, is
things (IoT) including social networks and e-commerce investigated based on the concepts of recommendation
systems, users always surf in the universe of data, in which systems (RS). Here, we apply the RS due to its ability to
users often do not know about who they are interacting with generate approximate trust value for unknown records based
and receiving data from. In such situations, the concept of on the available trustor–trustee relationships. Finally, we
trust plays an important role in managing these interactions discuss a realistic design model of the proposed items.
and developing a trustworthy environment for all providers,
users and the communities. However, generating trust From global standardization perspective, ITU-T Study
relationships among users is extremely hard due to Group (SG) 13 established the correspondence group on trust
diversified nature of the users and how each entity (CG-Trust) for preliminary work on trust standardization [8].
understands trust. In traditional forms of trust management The CG-Trust developed a technical report containing
systems, trust is computed based on the relationship among definition, use cases, functional classification as well as
end entities and behaviors in certain transactions as challenges, technical issues related to trust including overall
explained in [1], [2], [4-6]. Moreover, these systems use strategies of standardization for trust provisioning. As the
certain set of metrics like honesty, cooperativeness, lead group of trusted networking infrastructure, ITU-T SG13
community interest, reputation, certificate validity, successfully completed to publish the recommendation
length/frequency of the transaction and etc., to evaluate the Y.3052 on trust in March 2017 [9]. Recently Question 16/13
trustworthiness of end entities and then to find trust “Knowledge-centric trustworthy networking and services”
relationship among the trustors and the trustees. has focused on basic issues and key features on trust. Q16/13
is now mainly focusing on the development of core technical
However, trust on end entities is not always prominent but solutions for trust provisioning from ICT infrastructures and
the data receiving in form of various types. As an example, services. Q16/13 also plans to consider technology

978-92-61-24291-6/CFP1768P-ART © 2017 ITU – 147 – Kaleidoscope