A Multi-Constraint Path Optimization Scheme Based on Information Fusion in Software Defined Network

The existing multipath routing in Software Defined Network (SDN) is relatively blind and inefficient, and there is a lack of cooperation between the terminal and network sides, making it difficult to achieve dynamic adaptation of service requirements and network resources. To address these issues, we propose a multi-constraint path optimization scheme based on information fusion in SDN. The proposed scheme collects network topology and network state information on the network side and computes disjoint paths between end hosts. It uses the Fuzzy Analytic Hierarchy Process (FAHP) to calculate the weight coefficients of multiple constrained parameters and constructs a composite quality evaluation function for the paths to determine the priority of the disjoint paths. The SDN controller extracts the service attributes by analyzing the packet header and selects the optimal path for flow rule forwarding. Furthermore, the service attributes are fed back to the path composite quality evaluation function, and the path priority is dynamically adjusted to achieve dynamic adaptation between service requirements and network status. By continuously monitoring and analyzing the service attributes, the scheme can ensure optimal routing decisions in response to varying network conditions and evolving service demands. The experimental results demonstrated that the proposed scheme can effectively improve average throughput and link utilization while meeting the Quality of Service (QoS) requirements of various applications.