Trust-Based Parked Vehichular Fog For Catering Applications In 5G / Dyg Fatin Hamadah Bt Hj Awg Md Abdul Rahman.

Submitted in fulfillment of the requirements for the degree of Doctor of Philosophy in the Computer Network & Security School of Computing and Infromatics Universiti Teknologi Brunei.

ABSTRACT
The rise of 5G is envisioned to support applications with stringent latency such as virtual reality and autonomous vehicles. It is achievable with the help of fog computing as fog computing facilities are placed much closer to the end users. This results in reduced network band-width consumption as processes are being done locally, ultimately with low response time. Vehicles' capabilities are technologically advancing in terms of computation, networking and storage, however they are parked 96% of the time. This implies that the vehicles abundant resources can be a good source of computation in fog computing, with the existence and improved capability of vehicles' on-board units. The parked vehicles can be logically and physically clustered together to further enhance their computation capabilities and to ease the overall resource management. Although the vehicles are becoming highly connected, they are distributed and mobile in nature. Additionally, vehicles in general are heterogeneous such that they are made from different manufacturers and might have different platforms and security mechanisms. Vehicles have varying parking time which implies that they are dynamic and tempo- rary, resulting in the unpredictable behavior of the vehicles. Hence trust is needed to ensure seamless collaboration in the fog-based 5G environment. Trust is defined as the expectation that a device or system will faithfully behave in a particular manner to fulfill its intended purpose, and inevitably trust becomes a necessity in this diverse fog environment and is increasingly being used as part of decision making mechanisms.
There are numerous studies that have incorporated trust in vehicular networks involving fog computing where various metrics such as velocity, speed and direction are being used for trust evaluation. However, these metrics are not applicable in the parked vehicle scenario. Furthermore, no such study with regards to trust evaluation has been conducted in the parked vehicular fog. Thus, it is important to study the importance and application of trust in parked vehicle fog environment. In this the- sis, a framework for evaluating the trust of parked vehicle fog comput- ing facilities is proposed. The proposed framework forms a clustering concept based on trust value, where a physical vehicle cluster is subdivided into multiple logical trust clusters to allow tasks from end users to be distributed among the vehicular fogs. The framework comprises of a trust-based task scheduling algorithm which aims to improve the parked vehicle utilization and reduce task drop when there are client-generated tasks to be processed. Results demonstrate that the proposed framework is able to increase v-fog resource utilization and reduces percentage of task drop noticeably.
Additionally, the framework aims to cater to the latency requirement of applications in 5G. Realizing the benefits that virtualization technology can offer to the vehicle-based fog computing facilities, this thesis also studies the implication of trust-based allocation of virtual machines in vehicles. The proposed framework named EnTruVe is enhanced further where the results demonstrate that EnTruVe obtains the least number of VM migrations, it is the most energy efficient and has the highest utilization of v-fogs in comparison to other solutions

Doctorial Thesis (Degree) -Universiti Teknologi Brunei

Includes bibiliography (page 128-152)