Applying Social Event Data for the Management of Cellular Networks

TL;DR

This paper introduces a framework that leverages social event data to improve cellular network management by predicting events and their impact on network performance, aiding operations and maintenance tasks.

Contribution

It presents a novel system for automatic social data acquisition and processing, specifically designed for real-time cellular network management and performance prediction.

Findings

System successfully associates social events with network performance metrics

Framework improves network management efficiency in real deployment

Enhances prediction of network issues based on social event data

Abstract

Internet provides a growing variety of social data sources: calendars, event aggregators, social networks, browsers, etc. Also, the mechanisms to gather information from these sources, such as web services, semantic web and big data techniques have become more accessible and efficient. This allows a detailed prediction of the main expected events and their associated crowds. Due to the increasing requirements for service provision, particularly in urban areas, having information on those events would be extremely useful for Operations, Administration and Maintenance (OAM) tasks, since the social events largely affect the cellular network performance. Therefore, this paper presents a framework for the automatic acquisition and processing of social data, as well as their association with network elements (NEs) and their performance. The main functionalities of this system, which have been devised to directly work in real networks, are defined and developed. Different OAM applications of the proposed approach are analyzed and the system is evaluated in a real deployment.