Open Access

Christoph Frenzel

• Mobile communications are an integral part of modern society. People want to talk to their friends, play on-line games, or watch YouTube videos while they are on the go. This forces Mobile Network Operators (MNOs) to continuously upgrade and extend their mobile networks in order to keep up with the ever growing data rate requirements and provide the users with satisfactory Quality of Service (QoS). However, users are not willing to pay more for mobile communications leading to a decrease in the price per data unit. Due to this development, MNOs are faced with a serious pressure to reduce their Operational Expenditure (OPEX). The focus, thereby, lies on the reduction of manual efforts in network operations, i.e., the configuration and control of the Base Stations (BSs) such that the operational objectives by the MNO, defined as target values for network Key Performance Indicators (KPIs), are satisfied. Self-Organizing Network (SON) is a concept for automating the operation of mobileMobile communications are an integral part of modern society. People want to talk to their friends, play on-line games, or watch YouTube videos while they are on the go. This forces Mobile Network Operators (MNOs) to continuously upgrade and extend their mobile networks in order to keep up with the ever growing data rate requirements and provide the users with satisfactory Quality of Service (QoS). However, users are not willing to pay more for mobile communications leading to a decrease in the price per data unit. Due to this development, MNOs are faced with a serious pressure to reduce their Operational Expenditure (OPEX). The focus, thereby, lies on the reduction of manual efforts in network operations, i.e., the configuration and control of the Base Stations (BSs) such that the operational objectives by the MNO, defined as target values for network Key Performance Indicators (KPIs), are satisfied. Self-Organizing Network (SON) is a concept for automating the operation of mobile networks by relieving human operators from repetitive, low-level network operations tasks. In principle, a SON consists of a set of close-loop control functions, referred to as SON functions, which perform specific operations tasks. For instance, Coverage and Capacity Optimization (CCO) improves the network capacity and Mobility Robustness Optimization (MRO) optimizes the handover performance. The introduction of SON lifts the work of human operators from network operations to SON operations: instead of analyzing and configuring all the BSs in the network, the operational personnel has to configure and control the SON functions in order to adapt them to the network environment and specific operational objectives. This is due to the gap between the objectives and the low-level, technical configuration interface of the SON functions. As a result, MNOs ask for a further increase in the level of automation. This thesis introduces the Objective-Driven SON Operations (ODSO) concept that enables autonomic mobile network operations, i.e., the control of network operations through operational objectives on network KPIs. ODSO is built on top of SON and automates three manual tasks of SON operations: SON management that configures the SON functions, SON coordination that detects and resolves run-time conflicts between the concurrently running SON functions, and SON self-healing that detects and resolves failures in the network and the SON system. The specific contributions of this thesis are: 1. We analyze the specific requirements for autonomic operations of a mobile network and develop an architecture that extends an existing SON system with the ability to process the operational objectives of the MNO. The architecture integrates three components, one for each SON operations task, in order to reduce the manual control efforts. 2. We develop a generic decision making process that enables the ODSO components to autonomically control SON functions based on two types of information: a formalized representation of the operational objectives in an objective model, and a formalized representation of the task-specific technical knowledge in several technical models. On the foundation of Multiattribute Utility Theory (MAUT), the ODSO components are able to process these models in order to automatically operate the SON functions such that they optimize the network to satisfy the operational objectives. In this way, the MNO's objectives serve as a unified, high-level operations interface for SON. 3. We apply the generic decision making process to each of the three SON operations tasks which enables them to determine the best configuration of the SON functions, the best approach to resolve run-time conflicts, and the best recovery action to overcome failures with respect to the provided operational objectives. Thereby, we are able to account for the diverse characteristics of each task with our flexible approach. 4. We evaluate the feasibility and performance of the ODSO approach in a simulation of a realistic mobile network. As a result, we are able to show that the increased automation in SON operations reduces manual efforts and may also lift additional optimization potential in network operations. This leads to improved network performance compared to SON-based network operations.…