An Adaptive Forecasting Model for Slice Allocation in Softwarized Networks
Dyego H. L. Oliveira, Thelmo Pontes de Araujo, Rafael L. Gomes · IEEE Transactions on Network and Service Management · 2021
Internet Access Service (IAS) is a crucial tool for applications and system. An existing situation which affects the quality of IAS is the Elastic Demand for network resources. This situation arose the need of resource management evolution. The most promising approach for the evolution of IAS is the deployment of Softwarized Networks. Softwarized Networks allow the splitting of network resources into Slices, where each Slice can have the most suitable configuration. Within this context, this article presents a joint approach of an Adaptive Demand Forecasting model (ADF) and an Slice Allocation algorithm to define the most suitable slices structure based on the forecasting of the network resources demand. Experiments, using a real resource demand dataset, suggest that the proposal minimizes the error rate of foreseen values and it improves the network resource usage during the IAS deployment.