Year: 2018
Type: dissertation
Abstract: The work in this thesis proposes a number of energy efficient architectures of IoT networks. These proposed architectures are edge computing, Passive Optical Network (PON) and Peer to Peer (P2P) based architectures. \nA framework was introduced for virtualising edge computing assisted IoT. Two mixed integer linear programming (MILP) models and heuristics were developed to minimise the power consumption and to maximise the number of served IoT processing tasks. Further consideration was also given to the limited IoT processing capabilities and hence the potential of processing task blockage. Two placement scenarios were studied revealing that the optimal distribution of cloudlets achieved 38% power saving compared to placing the cloudlet in the gateway while gateway placement can save up to 47% of the power compared to the optimal placement but blocked 50% of the total IoT object requests. \nThe thesis also investigated the impact of PON deployment on the energy efficiency of IoT networks. A MILP model and a heuristic were developed to optimally minimise the power consumption of the proposed network. The results of this investigation showed that packing most of the VMs in OLT at a low traffic reduction percentage and placing them in relays at high traffic reduction rate saved power Also, the results revealed that utilising energy efficient PONs and serving heterogeneous VMs can save up to 19% of the total power. \nFinally, the thesis investigated a peer-to-peer (P2P) based architecture for IoT networks with fairness and incentives. It considered three VM placement scenarios and developed MILP models and heuristics to maximise the number of processing tasks served by VMs and to minimise the total power consumption of the proposed network. The results showed that the highest service rate was achieved by the hybrid scenario which consumes the highest amount of power compared to other scenarios. \n
Author: Zaineb T. Al-Azez
Language: English
Topic: IoT and Edge/Fog Computing
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2902788617