Resource optimization algorithm for 5G core network integrating NFV and SDN technologies

The growth in network demand has driven the development of new network technologies. However, traditional network architecture cannot meet the huge traffic of transportation and different business needs. To address this issue, a specific network service function chain is formed based on the network function virtualization. Dynamic resource awareness algorithms are introduced to construct an adaptive migration model based on network function virtualization. Based on the Multi-Armed Bandit (MAB) algorithm, a dynamic routing model based on MAB is constructed by using a greedy algorithm to search for random actions. When the nodes were 200 and 500, the migration costs of the adaptive migration model based on network function virtualization were 1000 and 3000, respectively. The average migration was 350 and 900 respectively, while destination nodes' average resource occupancy rates were 52 % and 58 %, respectively. When the path failure rates were 4 % and 20 %, the algorithm's safe path rates were 96.25 % and 92.75 %. For fixed and mobile nodes, the link load rate of the dynamic routing model based on the MAB algorithm was low and the load growth was relatively stable. This dynamic routing model's link delay is significantly less than the Dijkstra algorithm. These two models can maximize server resource utilization, reduce cost consumption, and achieve maximum overall benefits. • Propose a network function virtualization adaptive migration model for joint dynamic resource monitoring. • Multi-Armed Bandit routing algorithm with epsilon-greedy exploration dynamically selects optimal paths. • The combination of software defined networking and function virtualization enhances dynamic service functionality. • Software-defined networks can lower route delay. • Dynamic resource awareness algorithm can realize real-time tracking of CPU, memory and bandwidth usage.