The Internet's astounding success and growth can also lead to its gradual `ossification', hampering its ability to evolve and accommodate new services and applications. To address this problem, Software-Defined Networking (SDN) has been proposed as a way to 'program' networks and make it easier to deploy new protocols and new applications, as well as tune network performance. To-date, SDN techniques have, for the most part, targeted infrastructure-based networks. However, as self-organizing, infrastructure-less networks become more prevalent and ubiquitous, they will become integral parts of the Internets of the future, enabling a variety of applications such as vehicular communications, community services, healthcare delivery, emergency response, environmental monitoring, to name a few.
Motivated by this vision of the internet of the future, this project will develop Hybrid-SDN, or H-SDN, which will enable SDN in future internets consisting of infrastructure-based and infrastructure-less networks. To accomplish this goal, current SDN approaches, e.g., OpenFlow, which are inherently centralized, will be augmented so that they can operate in infrastructure-less, decentralized networked environments. The main deliverable of this project is the basic H-SDN prototype. Subsequently, we plan to (1) continue the H-SDN development and (2) develop efficient and flexible content delivery services in hybrid networks using H-SDN.
Broader Impact: By developing H-SDN, this research will likely have considerable broader impact as it will enable efficient content delivery in future internets. With content and computation moving to the cloud, efficient content delivery will be critical for content-- and service providers, as well as end users. This project also has a strong education component to this project including an exchange and co-supervision of junior researchers and graduate students between the two partner institutions.
