Improving Traffic Forecasting for 5G Core Network Scalability: A Machine Learning Approach

TLDR

5G will support a massive increase in traffic from IoT devices, relying on NFV and SDN to create flexible VNFs, and its new core architecture replaces point‑to‑point interfaces with producer/consumer communication, enabling dynamic scaling of VNFs based on traffic load. The study proposes a novel mechanism to scale 5G core network resources by anticipating traffic load changes through machine‑learning forecasting. Traffic load is forecasted using a neural network trained on real mobile traffic data, comparing LSTM‑based RNNs and DNNs. Simulations demonstrate that the forecast‑based scaling outperforms threshold‑based methods, reducing reaction latency and resource provisioning delay.

Abstract

5G is expected to provide network connectivity to not only classical devices (i.e., tablets, smartphones, etc.) but also to the IoT, which will drastically increase the traffic load carried over the network. 5G will mainly rely on NFV and SDN to build flexible and on-demand instances of functional networking entities via VNFs. Indeed, 3GPP is devising a new architecture for the core network, which replaces point-to-point interfaces used in 3G and 4G by producer/consumer-based communication among 5G core network functions, facilitating deployment over a virtual infrastructure. One big advantage of using VNFs is the possibility of dynamic scaling, depending on traffic load (i.e., instantiate new resources to VNFs when the traffic load increases and reduce the number of resources when the traffic load decreases). In this article, we propose a novel mechanism to scale 5G core network resources by anticipating traffic load changes through forecasting via ML techniques. The traffic load forecast is achieved by using and training a neural network on a real dataset of traffic arrival in a mobile network. Two techniques were used and compared: RNN, more specifically LSTM; and DNN. Simulation results show that the forecast-based scalability mechanism outperforms the threshold-based solutions, in terms of latency to react to traffic change, and delay to have new resources ready to be used by the VNF to react to traffic increase.