| dc.description.abstract | To meet the ever-growing network traffic demand, the underlying communication networks
need to be open and programmable. The Software Defined Networking (SDN) paradigm separates
the traditional data plane and network control plane, providing the ability to design, develop and
manage the communication networks in an efficient and scalable manner. SDN enables a high
degree of network programmability by providing high-level abstractions of the network devices
and interfaces to manipulate them. SDN has attracted significant attention in the research community
that produced a large body of work on programmable data plane and control frameworks.
However, the existing studies mainly focus on the wireline networks, while wireless networks have
received only limited attention.
Wireless traffic has significantly increased during the last couple of decades. Expansion of
wireless communication networks is setback by limited usable spectrum, varying channel conditions,
and signal fading. Envisioning to support re-configuration of wireless communication
networks, fast dynamic reconfiguration of radio devices, and robust network programmability is
necessitated. While network programmability is motivated by the SDN paradigm, Software Defined
Radio provides a platform to perform various signal processing functions in software rather
than hardware. However, to the best of our knowledge, there does not exist a framework to dynamically
program wireless devices across all levels of the protocol stack.
Accordingly, in this project, we build the foundations of a fully programmable SDN-enabled
framework that can achieve this goal. The proposed framework leverages the existing SDN approaches
for programmable data planes as well as the Software-Defined Radio paradigm. In particular,
we propose several key extensions to the SDN packet processing pipeline that enable different
per-flow behaviors at the physical layer. The contributions of this thesis include: (i) a set of different
physical layer profiles are presented as abstractions to the SDN application developer which
is used to set different Physical layer parameters for different traffic flow type; (ii) extensions of
SDN protocols (such as OpenFlow) to provide an abstraction for various physical layer profiles;
(iii) a proof-of-concept implementation and a set of use cases that demonstrate the benefits of our
approach. We also perform an extensive experimental study to evaluate the performance gains that
can be achieved by using fully programmable wireless devices. | en |
