VidSDN is a four years project funded by Science Foundation Ireland (SFI). The ultimate goal of VidSDN is to create an application-aware software defined network (SDN) so that application providers can cooperate with network and cloud operators to enhance the application performance and at the same time optimize the network and cloud resource usage through Software Defined Networking technologies. VidSDN contains two research strands.

Strand 1: Application-aware SDN

VidSDN Architecture 1
Application providers (e.g. VoIP and Video Conferencing companies) need to dynamically change the network topology and re-allocate resources according to various factors, e.g. user number, user location, user capabilities, network bandwidth, fault tolerant requirements, and load balancing requirements. The network contains more and more resources/middleboxes, e.g. caching, computing, redundancy elimination, compression, encryption. Current situation is that the two parties work separately. From the applications’ perspective, the network is transparent and the network conditions fluctuate unpredictably. Therefore, current application providers try to work around the network rather than work with the network. From the network’s perspective, the network resources are shared by all the applications. It is difficult to guarantee quality of services for each application. There are also no convenient approaches to expose network resources, e.g. middleboxes, to applications to further increase revenue.

This strand looks for a solution that can enable the two parties work together and benefit both of them. It focuses on designing a generic SDN framework to enable application awareness in the network and cloud. It will enable fine-grained control of application traffics, boost creation of new application services using network resources, and also generate new network services for network and cloud operators.

Strand 2: Video Delivery over ICN

VidSDN Architecture 2
The current Internet architecture was created in 1970s and was originally aimed for sharing expensive resources. This leads to a host centric communication model, i.e. one host providing resources and the other host communicates with the server host to access the resources. As the tremendous growth of applications and services, the original design of the Internet architecture cannot fulfil the new requirements, e.g. requirements in mobility, security, and scalability. A novel Internet architecture called Information Centric Networking (ICN) was proposed to address these problems, which has become one of the most active research areas in computer networking communities. The ICN architecture is based on the observation that the current Internet is primarily used for content distribution instead of communication between two hosts and users are more interested in what than where. In ICN, the content (or information) are named and routed directly in the network. There is no notion of host at the lowest level.

This strand focuses on applying ICN technologies into video distribution scenarios to enhance the application performance and at the same time refine the current ICN design. It will develop an ICN based video distribution solution through designing video content naming, routing, and caching schemes. The interactions between network resources and application requirements will be considered in the performance optimization. The strand will also investigate issues in deploying ICN over the existing Internet.

Dr. Yuansong Qiao
Yuhang Ye
Dr. Brian Lee
Dr. Ronan Flynn