Big data-driven optimization for performance management in mobile networks

Abstract

Humanity, since its inception, has been interested in the materialization of knowledge. Various ancient cultures generated a lot of information through their writing systems. The beginning of the increase of information could date back to 1880 when a census performed in the United States of America took 8 years to be tabulated. In the 1930s the demographic growth exacerbated this increase of data. Already in 1940, libraries had collected a large amount of writing and it is in this decade when scientists begin to use the term “information explosion”. The term first appears in the Lawton (Oklahoma) Constitution newspaper in 1941. Currently, it can be said that we live in the age of big data. Exabytes of data are generated every day; therefore, the term big data has become one of the most important concepts in information systems. Big data refer to large amounts of data on a large scale that exceeds the capacity of conventional software to be captured, processed, and stored in a reasonable time. As a general criterion, most experts consider big data to be the largest volume of data, the variety of formats and sources from which it comes, the immense speed at which it is generated, the veracity of its content, and the value of the information extracted/processed. Faced with this reality, several questions arise: How to manipulate this large amount of data? How to obtain important results to gain knowledge from this data? Therefore, the need to create a connecting bridge between big data and wisdom is evident. People, machines, applications, and other elements that make up a complex and constantly evolving ecosystem are involved in this process. Each project presents different peculiarities in the development of an framework based on big data. This, in turn, makes the landscape more complex for the designer since multiple options can be selected for the same purpose. In this work, we focus on an framework for processing mobile network performance management data. In mobile networks, one of the fundamental areas is planning and optimization. This area analyzes the key performance indicators to evaluate the behavior of the network. These indicators are calculated from the raw data sent by the different network elements. The network administration teams, which receive these raw data and process them, use systems that are no longer adequate enough due to the great growth of networks and the emergence of new technologies such as 5G and 6G that also include equipment from the Internet of things. For the aforementioned reasons, we propose in this work a big data framework for processing mobile network performance management data. We have tested our proposal using performance files from real networks. All the processing carried out on the raw data with XML format is detailed and the solution is evaluated in the ingestion and reporting components. This study can help telecommunications vendors to have a reference big data framework to face the current and future challenges in the performance management in mobile networks. For instance, to reduce the processing time data for decisions in many of the activities involved in the daily operation and future network planning.