A Fog Computing Approach for Cognitive, Reliable and Trusted Distributed Systems

In the Internet of Things era, a big volume of data is generated/gathered every second from billions of connected devices. The current network paradigm, which relies on centralised data centres (a.k.a. Cloud computing), becomes an impractical solution for IoT data storing and processing due to the long distance between the data source (e.g., sensors) and designated data centres. It worth noting that the long distance in this context refers to the physical path and time interval of when data is generated and when it get processed. To explain more, by the time the data reaches a far data centre, the importance of the data can be depreciated. Therefore, the network topologies have evolved to permit data processing and storage at the edge of the network, introducing what so-called fog Computing. The later will obviously lead to improvements in quality of service via processing and responding quickly and efficiently to varieties of data processing requests. Although fog computing is recognized as a promising computing paradigm, it suffers from challenging issues that involve: i) concrete adoption and management of fogs for decentralized data processing. ii) resources allocation in both cloud and fog layers. iii) having a sustainable performance since fog have a limited capacity in comparison with cloud. iv) having a secure and trusted networking environment for fogs to share resources and exchange data securely and efficiently. Hence, the thesis focus is on having a stable performance for fog nodes by enhancing resources management and allocation, along with safety procedures, to aid the IoT-services delivery and cloud computing in the ever growing industry of smart things. The main aspects related to the performance stability of fog computing involves the development of cognitive fog nodes that aim at provide fast and reliable services, efficient resources managements, and trusted networking, and hence ensure the best Quality of Experience, Quality of Service and Quality of Protection to end-users. Therefore the contribution of this thesis in brief is a novel Fog Resource manAgeMEnt Scheme (FRAMES) which has been proposed to crystallise fog distribution and resource management with an appropriate service's loads distribution and allocation based on the Fog-2-Fog coordination. Also, a novel COMputIng Trust manageMENT (COMITMENT) which is a software-based approach that is responsible for providing a secure and trusted environment for fog nodes to share their resources and exchange data packets. Both FRAMES and COMITMENT are encapsulated in the proposed Cognitive Fog (CF) computing which aims at making fog able to not only act on the data but also interpret the gathered data in a way that mimics the process of cognition in the human mind. Hence, FRAMES provide CF with elastic resource managements for load balancing and resolving congestion, while the COMITMENT employ trust and recommendations models to avoid malicious fog nodes in the Fog-2-Fog coordination environment. The proposed algorithms for FRAMES and COMITMENT have outperformed the competitive benchmark algorithms, namely Random Walks Offloading (RWO) and Nearest Fog Offloading (NFO) in the experiments to verify the validity and performance. The experiments were conducted on the performance (in terms of latency), load balancing among fog nodes and fogs trustworthiness along with detecting malicious events and attacks in the Fog-2-Fog environment. The performance of the proposed FRAMES's offloading algorithms has the lowest run-time (i.e., latency) against the benchmark algorithms (RWO and NFO) for processing equal-number of packets. Also, COMITMENT's algorithms were able to detect the collaboration requests whether they are secure, malicious or anonymous. The proposed work shows potential in achieving a sustainable fog networking paradigm and highlights significant benefits of fog computing in the computing ecosystem