A Software-Defined Self-Adaptive Wireless Network Design

In recent years network design has grown to accommodate emerging areas of application such as smart cities. The "smart city" concept combines the information and communications technology; where heterogeneous end devices with different capabilities and resource requirements coexist. Thousands of these devices form a wireless mesh network (WMN), wireless sensor network (WSN), or an Internet of Things (IoT) to enable efficient information collection from the heterogeneous devices. The gathered information is then analyzed to realize appropriate and timely decisions for various city operations and services [46]. The "smart city" applications include smart traffic lights, streetlights, surveillance, emergency services, and environmental monitoring. The networks in smart cities must react to communication link and node failures in a matter of milliseconds and ideally be able to predict future states and anticipate the demands of applications, services, and network environments. For instance, city-scale video surveillance applications demand high bandwidth to guarantee timely delivery of critical events to interested observation centers. To enable these networks to reach their intended potential, new models must be developed that can enable the reliable, efficient, and adaptive behavior of the networks; these are properties that cannot be supported by the current generation of wireless network designs that are based on legacy networking stack architecture. The proposed research will fill the gap between existing designs of wireless networks and demands of emerging wireless network applications such as in smart cities. Software-defined networking (SDN) is on track to become the standard network architecture; thus, my design of next-generation wireless networks will be structured to fit SDN paradigm. The SDN architecture is composed of management, control, and infrastructure (data) planes, where network policy is defined at the management plane. The control layer then translates policies into network configuration rules for infrastructure layer devices, according to a global network view. SDN further supports network programmability and implementations that are portable and extensible. The proposed research will design new algorithms, protocols, and architectures to enable reliable network operations. The control plane design will propose a new concept of the possible separation to a fast and a slow changing network state behavior to support efficiency and agility. The adaptive network operation will be based on a novel idea of combining network data flow context and analytics, which honors constraints from different SDN planes. I will produce a programmable, reliable, efficient, agile, and self-adaptive wireless network design targeted for, but not limited to, "smart city" applications, such as smart streetlights, traffic lights, and surveillance.