Towards Softwarized Drone Networks

Abstract

Drones, or Unmanned Aerial Vehicles (UAVs), are considered essential tools in search and rescue, disaster relief, remote sensing, aerial surveillance and security. Drone-assisted communication networks are gaining considerable attention as a cost-effective and flexible network infrastructure offering new capabilities and opportunities. In addition to enabling connectivity for complex multi-drone tasks, drone networks can be deployed to facilitate connectivity in remote areas and extreme environments and supplement and extend the coverage of mobile networks in response to variable demands. However, utilizing such flexibility requires dealing with the inherent dynamics of drone networks, characterized by a high level of mobility and limited resources. Network softwarization using Software Defined Networking (SDN) and Network Functions Virtualization (NFV) enables flexible and adaptive control and reconfigurability in drone networks through centralized programmability and virtualized network functions.

In this thesis, we investigate drone network softwarization by identifying potential gains from the flexibility offered by softwarization not investigated previously in the context of drone networks. To enable the utilization of SDN and NFV in drone networks, we propose and describe an architecture for softwarized drone networks. Furthermore, we address an important challenge relating to SDN control, a key element in SDN architectures, allowing the programmability of the network through an interface between logically centralized controllers and the programmable network nodes. To adapt to the network mobility and connectivity constraints, we propose schemes for deploying and assigning SDN controllers embedded in drones, allowing for continuous operation of control functions with changing network topology and possible unavailability of ground infrastructure. We also utilize the flexibility offered by NFV. New deployment and orchestration schemes are needed to efficiently deploy and manage drone networks defined by Virtual Network Functions (VNFs) implementing task and network functionalities. To this end, we describe the applicable use cases that benefit from this flexibility and propose schemes that efficiently deploy and manage NFV-based drone networks.