ABSTRACT
The evolution of mobile technology has been driven by the ever-increasing demands for high data rates and diverse applications. This has significant implications for the network infrastructure providers or mobile operators. The huge imbalance in the growth rates of data throughputs and operators’ revenue implies that both capital expenditure (CAPEX) and operation expenditure (OPEX) need to be reduced. Furthermore, the challenges on the technical front grow rigid due to the complexity and scale of the modern mobile communication system. Operators pursuing approaches of combining heterogeneous access technologies to boost their network services adds another layer of complexity in network management. Therefore, it is believed that the traditional (mainly human-controlled) network management paradigm necessitates a shift toward a self-organization and self-optimization system, which assists in enacting the goal of reduced OPEX. The implication of self-x network management is to enable the networks to organize and optimize their parameters by themselves and to minimize human intervention [1]. Most of the proposals to realize self-x network management are at large inspired by the biological systems that exhibit autonomic behavior, such as 187self-healing, self-management, and so on. This entails that for networks to fully implement the self-x vision, the following autonomic principles need to be implemented: (1) the ability to translate business goals into low-level network configurations; (2) the in-time sensing of contextual changes in the networks and the timely reporting of them to proper network segment(s); (3) the implementation of an optimal control behavior upon sensing contextual changes, which ensures that the system’s functionality adapts to meet the requirements of the changing environment; (4) the capability to observe the impact of its extended control strategy and to learn to converge with an optimal strategy.
