Abstract
The University of Central Florida invention is a novel Speculative Software-Defined Networking (SDN) framework that incorporates reinforcement learning (RL) to predict the arrival of flows that may not have been seen before. The invention shows that the RL agents can learn and speculatively install unseen flow rules to avoid additional control latency due to the reactive installation of flow rules. The UCF design helps respond to application dynamics. It makes reactive SDN a strong candidate for responding to the needs of emerging low-latency applications, such as augmented and virtual reality.
Partnering Opportunity: The research team is looking for partners to develop the technology further for commercialization.
Benefit
Efficiency and effectiveness in terms of finding the best group of flow rules in SDNs for low-latency applicationsReward function increases the efficacy of the RL agents' capability in predicting the best set of flows to install in the switch's flow tableCan help to significantly reduce the miss rate at Reactive SDN switches, which is crucial for low-latency applications such as online gaming, augmented or virtual realityMarket Application
Low-latency applications such as online gaming and augmented reality (AR)/virtual reality (VR) Publications
RL-Based
Speculative Installation of Unseen Flows in SDNs for Low-Latency Applications,
Interactive Session 5: Network Optimization II, IEEE International Conference
on Machine Learning for Communication and Networking (ICMLCN), Stockholm,
Sweden, May 5-8, 2024.
Brochure
