Abstract
One of the models in the literature for modeling the behavior of the brain is the Bayesian attractor model, which is a kind of machine-learning algorithm. According to this model, the brain assigns stochastic variables to possible decisions (attractors) and chooses one of them when enough evidence is collected from sensory systems to achieve a confidence level high enough to make a decision. In this paper, we introduce a software defined networking (SDN) application based on a brain-inspired Bayesian attractor model for identification of the current traffic pattern for the supervision and automation of Internet of things (IoT) networks that exhibit a limited number of traffic patterns. In a real SDN testbed, we demonstrate that our SDN application can identify the traffic patterns using a limited set of fluctuating network statistics of edge link utilization. Moreover, we show that our application can improve core link utilization and the power efficiency of IoT networks by immediately applying a pre-calculated network configuration optimized by traffic engineering with network slicing for the identified pattern.
Highlights
In conventional networks, networking devices are basically composed of a data plane, which handles the processing, modifying, and forwarding of packets, and a control plane, which decides the port when forwarding the data
To prevent congestion and increase the quality of service (QoS) in Internet of things (IoT) networks like the surveillance networks that exhibit a limited number of traffic matrices, we propose an Software defined networking (SDN)-based traffic engineering framework that uses a different methodology for estimating the traffic matrix
We investigated the feasibility of applying a brain-inspired Bayesian attractor model (BAM) and showed that it can identify the traffic in a reasonable time without such requirements
Summary
In conventional networks, networking devices are basically composed of a data plane, which handles the processing, modifying, and forwarding of packets, and a control plane, which decides the port when forwarding the data. The solutions proposed by ML can be applied to all network devices in the network in real-time by the SDN controller [2]. There are many proposals in the literature for estimating the traffic matrix, but they have important trade-offs like training a neural network for a long time, capturing traffic from the interfaces, high bandwidth or CPU usage etc. To prevent congestion and increase the QoS in IoT networks like the surveillance networks that exhibit a limited number of traffic matrices, we propose an SDN-based traffic engineering framework that uses a different methodology for estimating the traffic matrix. The experiments reveal that our SDN framework can correctly identify a changing traffic pattern by BAM and increase the QoS and the energy efficiency by applying an optimized configuration employing network slicing and traffic engineering.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
