DYNAMIC SOFTWARE DEFINED NETWORKING PLANE SCALING METHOD AND SYSTEM BASED ON LEARNING CURVES IN INTERNET OF THINGS

摘要

The present invention relates to a dynamic software-defined networking plane scaling method based on a learning curve in an IoT network environment, and more specifically, as a dynamic software-defined networking plane scaling method, The dynamic software-defined networking plane includes an available plane composed of a plurality of unit planes, but (1) requesting allocation of an available plane based on the number of data packets currently occurring; (2) predicting the number of unit planes required in the data plane using a prediction algorithm based on a learning curve according to the request of step (1); And (3) assigning a unit plane from an available plane to a data plane according to the number of required unit planes in the data plane predicted in step (2), and allocating the remaining unit planes as a control plane, dynamically defining a software-defined networking plane. It comprises a step of scaling the feature of the configuration. In addition, the present invention relates to a dynamic software defined networking plane scaling system 10 based on a learning curve in an IoT network environment, and more specifically, as a dynamic software defined networking plane scaling system 10, the dynamic The software-defined networking plane includes an available plane composed of a plurality of unit planes, but an available plane request unit 100 requesting allocation of an available plane based on the number of data packets currently occurring; A prediction unit 200 for predicting the number of unit planes required in the data plane using a prediction algorithm based on a learning curve at the request of the available plane request unit 100; And assigning a unit plane from an available plane to a data plane according to the number of required unit planes in the data plane predicted by the prediction unit 200, and assigning the remaining unit planes to the control plane, dynamically software-defined networking planes. It comprises a scaling unit 300 for scaling is characterized by its configuration. According to the dynamic software-defined networking plane scaling method and system 10 based on the learning curve in the IoT network environment proposed in the present invention proposed in the present invention, the available plane is composed of a plurality of unit planes, The number of data planes and control planes can be efficiently adjusted by reflecting the properties of changing data packets. In addition, according to the dynamic software-defined networking plane scaling method and system based on a learning curve in the IoT network environment proposed in the present invention, data required by using a linear regression algorithm as a prediction algorithm based on the learning curve By quickly predicting the number of planes, a unit plane may be allocated from an available plane to a data plane according to the predicted result, and the remaining unit planes may be allocated as a control plane. In addition, according to the dynamic software-defined networking plane scaling method and system based on the learning curve in the IoT network environment proposed by the present invention, the properties of data packets changing in real time from the IoT or smart dust network environment can be determined. The plane can be dynamically set by reflecting it, so that the plane can be adjusted more efficiently, and the number of planes can be appropriately determined even in the case of a disconnection or a sudden increase in the amount of data packets.