| Summary: | The establishment of smart city database requires Internet of things technology. With the continuous expansion of the scale of the Internet of things data center, the energy consumption of the data center is increasing, which greatly limits the development of the Internet of things data center. Energy consumption of Internet of things network data center in China's smart city has increased year by year, and data center network energy conservation has become a current research hotspot. The traditional network nodes are distributed and centralized innovatively to complete centralized control, and the controller completes the collection of the whole network information, the maintenance of network status, the distribution of flow entry, etc. Firstly, the energy-saving model of digital center network and the model data center network traffic prediction algorithm based on the principle of high accuracy traffic prediction are constructed, and an energy-saving multi-layer virtual traffic scheduling algorithm is proposed. Secondly, the two algorithms are fused, and finally an empirical study is carried out. The results show that in Random mode, the energy consumption ratio of energy-efficient multi-layer virtual-software defined networking (EMV-SDN) is the lowest, and the maximum reduction of energy consumption ratio reaches 7.8% compared with equal-cost multi-path (ECMP) algorithm. In Staggered mode and Stride mode, the energy consumption ratio of EMV-SDN algorithm is the lowest. Compared with the actual data flow, the prediction result of K-means-support vector machine (KM-SVM) algorithm is closer to the actual result, and the maximum error between the predicted value and the actual value of KM-SVM algorithm is 1.2 Gbps. However, the maximum error of balanced iterative reducing and clustering using hierarchies-support vector machine (B-SVM) algorithm reaches 3.1 Gbps; the prediction accuracy of KM-SVM algorithm is always higher than that of B-SVM algorithm in both discontinuous data flow and continuous actual data flow; the accuracy of KM-SVM algorithm in different experiments is high; among the energy consumption ratios of the combined algorithm, the energy consumption ratio of EMV-SDN algorithm is the lowest under the three communication modes, and the performance of the algorithm constructed in this study is higher than that of other algorithms in simulation operation; when the traffic prediction algorithm is combined with the virtual topology energy conservation control algorithm, the network structure change is reduced and the network stability is increased in this study. In the delay comparison of EMV-SDN algorithm, ECMP algorithm, and Dijkstra algorithm in the three communication modes, the EMV-SDN algorithm has the best performance. The results of this study provide an improvement direction for the development of Internet of Things technology and the construction of smart cities.
|
|---|