Software Defined Networking Architecture with Adaptive Scheduling Concept In Internet of Thing Environment

Abstract

With the development of the Internet of Things many problems have been extended For example the Internet of Things is widely deployed in multiple different sub-networks and the use of different wireless communication methods is extremely heterogeneous How to effective integration these sub-networks into a seamless communication platform; Man-aging geographically dispersed heterogeneous networks in a dynamic environment has been a major research in the past few years One of the used methods is the SDN (Software Defined Networking Architecture) which aggregates topological information through a Centralized View to Calculate the best route flow for the entire domain but deploying SDN in IoT is different from in tradition environment We have found that past research attempts to build a SDN method on a Spontaneous Network with infrastructure-less and a highly heterogeneous network enable consider application-level QoS capabilities on the basis of multi-hop heterogeneous nodes dispatching packets and proposes a priority mechanism to strengthen the concept of specific applications occupying network resources However in this study although consideration was given to assigning priority to network resource occupation time on a single path it did not consider the possibility of having the opportunity to choose other paths in the network topology If only time schedule is achieved and the path is spontaneously generated by the local end it is very likely that only a single application flow can achieve better QoS and it is compressed to the possibility that other applications can choose other sub-optimal routes This article will continue the architecture of the previous research and add the concept of flow routing scheduling use Genetic Algorithm to help SDN controller solve hte NP-Complete problem of multi-restricted path selection in which the Formal Method is used to reasonably estimate the load of network resources and the mechanism of allocating network resource occupation time is activated only on Overloaded links enabling our SDN architecture to self-adjust to optimal scheduling strategies in dense and sparse network topology environments

Date of Award	2020
Original language	English
Supervisor	Chuan-Ching Sue (Supervisor)