Optimizing SDN resource allocation using fuzzy logic and VM mapping technique

Network virtualization refers to running multiple heterogeneous architectures simultaneously on a shared substrate. One major challenge in network virtualization is virtual network mapping to a substrate network. Effective management of substrate network resources during mapping is crucial for optimizing these resources. Software defined networking (SDN) is a novel technology that separates control logic from data, and is utilized in network virtualization to improve resource management. In this study, we leverage SDN for network virtualization, allowing for dynamic management of substrate resources to achieve optimal mapping. We use a fuzzy Markov model and time slots to optimize response time and improve service quality. The model predicts the appropriate size for sending packets in the next slot and ensures successful mapping, resulting in the most efficient use of network resources. This approach reduces costs and delays, increases acceptance rates, improves quality of service, and ultimately enhances productivity. We use NS2 and Mininet simulators with acceptance rate and latency as evaluation criteria to evaluate our approach. The results indicate a significant improvement over previous research regarding request acceptance rate and time. Specifically, the switch link-resources and acceptance rates show a 7.12% and 9.34% improvement rate, respectively, when compared to SDN-nSR and SDN-NV techniques.