A Metaheuristic Multi-Constraint Routing Mechanism for Traffic Scheduling in Software-Defined Networking

Abstract

Due to the rapid development of the Internet more and more network services are beginning to emerge such as multimedia transmission network streaming video conferencing etc These types of network services require different levels of Quality of Service (QoS) requirements However the Best Effort delivery principle of the traditional Internet Protocol (IP) cannot satisfy these needs Therefore providing appropriate routing mechanism for these network applications has become an important issue In order to achieve these requirements two main methods have been proposed in the traditional network The Integrated Service (IntServ) provides the bandwidth to the corresponding route by using the Resource Reservation Protocol (RSVP); Differentiated services (DiffServ) use the method of packet classification to achieve different service quality requirements However due to the influence of decentralized architecture in the traditional network neither of these methods has an effective subordinate and achieves fast and consistent routing planning In recent decades scholars have proposed an emerging architecture which could refer to Software Defined Networking (SDN) Under this architecture the network is divided into Control Plane and Data Plane Authority of management is delegated to the control layer's controller which makes the data network more flexible easier to operate and manage and better able to cope with application development And with its central centralized control mode we can instantly monitor traffic and re-adjust the network very quickly to achieve specific goals Therefore a management mechanism developed under software-defined network can achieve better routing planning However past research focused only on the optimization of service of quality and did not take into account the memory size limitations of network devices under the SDN architecture Based on the above issues this study proposes a metaheuristic routing mechanism to achieve QoS requirements of different flows while avoiding memory overflow problem in OpenFlow switch Finally the simulation results show that the research mechanism can not only achieve network application traffic for different service quality requirements but also avoid overflow of network device memory

Date of Award	2018 Aug 31
Original language	English
Supervisor	Hui-Tang Lin (Supervisor)