Cellular Uplink Bandwidth Prediction Based on Radio Measurements

Abstract

In 4G networks, the emergence of machine communications such as connected vehicles increases the high demand of uplink transmissions, thus, degrading the quality of service per user equipment. Enforcing quality-of-service in such cellular network is challenging, as radio phenomenon, as well as user (and their devices) mobility and dynamics, are uncontrolled. To solve this issue, estimating what the quality of transmissions will be in a short future for a connected user is essential. For that purpose, we argue that radio metrics are key features whose evolutions can help predicting the bandwidth that the considered connections can take advantage of in the following hundreds of milliseconds. The paper then describes how a 4G testbed has been deployed in order to study the correlation between radio noise and throughput in uplink transmissions. Based on radio measurements, the main supervised machine learning algorithms are used, such as Random Forest and Support Vector Machine to predict the uplink received bandwidth. For a specific user service, we are able to predict the end-to-end received bandwidth, i.e. the amount of received data on the server side during a specific period at a very low scale of 100 ms. Results also prove that uplink bandwidth predictions are less accurate compared to bandwidth prediction for downlink based on radio measurements.

Connection Type Identification and Uplink Speed Estimation of Malware Infected Hosts

“…Additionally, the uplink speed of the network is widely used to estimate the pattern of data transmission [10], and to calculate the scale of malicious access to abuse the Internet resources [11], etc. As a reference, uplink speed is analyzed in alternative conventional works [8], [9], [12], [13], [14], [15], [16], [17], [18], [19]. In detail, the work [8] showed the value of combining information from several sources to spatially distinguish significant sites and roughly locate the user utilizing network artifacts, like IP addresses, which are gathered from mobile applications.…”

“…Work [16] worked on multiple linear regression methods to predict hourly downlink speed. Work [17] investigated the uplink bandwidth prediction in the cellular networks using the machine learning method, which focused on quality of service based on a signal-noise ratio. Work [18] focused on the cellular uplink speed prediction between IoT devices in vehicle connection.…”