Towards Scalable and Efficient Software-defined Networking for Internet of Things

Project Description

Optimal Planning for Distributed Control Plane

To enable the distributed control plane for SDN in IoTs, we need to address several emerging challenges. First, we need to develop controller planning for efficient and scalable multi-controller placement. Second, we need to design fast control traffic balancing and inter-controller synchronization to facilitate online and adaptive traffic engineering and to localize all decisions to each controller for minimizing control plane response time, respectively.

Scalable and Adaptive Software-Defined Traffic Engineering

To support scalable and adaptive software-defined traffic engineering upon a distributed control plane, we aim to simultaneously meet the strict QoS requirements of real-time traffic and maximize the throughput of best effort flows, while providing highly scalable optimized performance.

SDN-enhanced Fog Computing for Mobile IoTs

Fog computing extends the cloud-based Internet by introducing an intermediate fog layer between devices and cloud, aiming at providing the smooth, low-latency computing service to devices. Despite its great advantage, fog computing faces fundamental challenges when the things, e.g., wearable sensors, smart phones, and connected vehicles, become increasingly mobile. To counter this problem, first, it is highly desirable to intelligently connect the mobile thing to a new fog server in its vicinity, which has sufficient computing, memory and wireless bandwidth resources. Second, besides selecting optimal fog sever, connecting to a new server requires the efficient live virtual machine migration, which is a process of relocating the virtual machine of the mobile thing from its previous hosting fog server to the new one, with the minimum service disruption time.

Testbed Development and Experimental Plan

We view validation of the proposed technology as a most important facet of this proposal. To validate the proposed SDN solutions for IoTs, we will leverage our expertise and prior experience in SDRs to develop a new experimental facility (initially housed at Wichita State University, and then to be replicated at GaTech) as a proof of concept.