Optimisation of Wireless Disaster Telecommunication Network based on Network Functions Virtualisation under special consideration of Energy Consumption

Abstract

A working communication network is a key element in saving human lives after a disaster event. However, analysis of past disaster events shows significant damage to common communication network infrastructures such as the mobile network or landline. Consequently, a new communication network (Emergency Communication Network (ECN)) must be established immediately after the disaster to support rescue operations (e.g., civil protection, police, and volunteers) and to enable communication between the affected people. The established network is subject to several requirements. These requirements can be divided into infrastructure and service requirements. Infrastructure requirements include rapid deployment of the communication network, complete coverage of the disaster area, access for most of the terminals in use, failure safety of the network infrastructure and QoS support for the provided services. Service requirements include the scalability and failure safety of the provided services, as well as the quick integration of new services. This research work aims to design a communication network that meets the requirements in the event of a disaster. For this purpose, a new network architecture is proposed. The proposed network architecture improves the performance of a wireless disaster network (Wireless Mesh Network (WMN)) through integration with Network Functions Virtualisation (NFV). This improvement is achieved by (first) eliminating dedicated hardware equipment (e.g., vehicles with heavy proprietary communication middleboxes, servers, satellite, and TETRA phones) and replacing them with conventional wireless routers, (second) introducing NFV to address the unresolved requirements of service scalability and rapid integration of new services, and (third) optimising energy consumption and resource use to solve the network sustainability problem. This thesis has three significant contributions. Firstly, it investigates the routing in NFV optimised WMN for disasters and compares the performance of different routing protocols. It proposes a two-layer architecture that combines the advantages of both layer 2 and layer 3 routing and enables routing in large networks. Secondly, it addresses the throughput in WMN and proposes a new channel allocation scheme. The proposed channel assignment method is cluster-based and addresses issues such as the optimal cluster size or the maximum number of wireless interfaces. Previous works have not investigated these issues. This thesis's third contribution and focus are to develop algorithms for energy-efficient placement of VNFs in WMN. Since the devices forming the network in case of a disaster are often battery-powered due to damage to the power grid, the choice of a location for a VNF directly impacts the energy consumption and thus on the network's lifetime. Besides the mathematical formulation of the optimisation problem as multi-objective optimisation, the proposed algorithms are implemented, and their performance is compared in different scenarios.