MEMBRANE: ModEling Mobile BRoAdband NEtworks

Mobile broadband (MBB) networks underpin numerous vital operations of the society and are arguably becoming the most important piece of the communications infrastructure. The use of MBB networks has exploded over the last few years due to the popularity of mobile devices, combined with the availability of high-capacity 3G/4G mobile networks. Given the increasing importance of MBB networks, there is a strong need for a better understanding of the fundamental characteristics of MBB networks and their relationship with the performance of popular applications, such as mobile video. This is crucial for different stakeholders, from operators to application developers, to improve the quality of their services and products. The overarching goal of the MEMBRANE project is to address this need by developing reliable models that can accurately capture the relationship between the performance and reliability metrics and the underlying network parameters that affect them. These models reveal the key factors that are required to accurately describe and predict the performance and reliability of MBB networks. To achieve this goal, MEMBRANE follows a measurement-driven modelling approach. As the first step, we run an extensive measurement campaign to collect data from operational MBB networks in order to understand their performance and reliability. Second, we leverage machine-learning approaches to derive models that enable us to better understand the relationship between MBB network characteristics and the performance of the applications experienced by the end users. Over the last few years, we have extensively collected data from open platforms such as MONROE as well as leverage crowdsourced datasets. Our focus has been to assess how the MBB networks perform, e.g., from the perspective of performance and reliability such as availability, coverage, and network Quality of Service (QoS). We observe that mobile networks, especially in Scandinavia, provides good coverage and very high data rates in the urban areas. Considering the MBB performance is evaluated by not only how the network behaves in general but also how specific services are provisioned by the network, we extended this work to focus on Quality of Experience (QoE) for two key applications: web services and video streaming. For these applications, we showed that mobility is the most critical factor that impacts the QoE of the users. Recently, we also considered emerging application such as 360-degree video streaming over mobile networks and designed a framework that enables us to evaluate the performance both in operational networks as well as 5G testbeds. Finally, we extended our focus also on the Internet of things (IoT) direction and started the evaluation of Narrow Band IoT (NB-IoT) performance, both in terms of coverage and energy efficiency, in Norway.

Models developed in MEMBRANE revealed the key factors that are required to accurately describe and predict the performance and reliability of MBB networks. More specifically, we have considered performance on the current 3G/4G infrastructures, and we have proposed solutions to provide a better user experience for the end users. We further provided methods to improve the user experience for the 5G technologies. An important aspect of 5G is to support massive Internet of Things (IoT) applications and MEMBRANE extended the performance analysis also in cellular IoT domain with Narrow Band IoT analysis. These works resulted in 20 conference/journal articles that are published in well respected venues in the communication and networking fields. The results are further communicated to different stakeholders including regulators, operators, businesses and application developers to be considered to improve services and products.

Mobile broadband (MBB) networks underpin numerous vital operations of the society and are arguably becoming the most important piece of the communications infrastructure. The use of MBB networks has exploded over the last few years due to the popularity of mobile devices, combined with the availability of high-capacity 3G/4G mobile networks. More than half of this mobile traffic is generated by mobile video. Given the increasing importance of MBB networks, there is a strong need for a better understanding of the fundamental characteristics of MBB networks and their relationship with the performance of popular applications, such as mobile video. The overarching goal of the MEMBRANE project is to address this need by developing reliable models that can accurately capture the relationship between the performance and reliability parameters and the underlying network factors that affect them. These models will thus reveal the key factors that are required to accurately describe and predict the performance and reliability of MBB networks in general and mobile video in particular. To achieve this goal, we will follow a real-world measurement-driven modeling approach. As the first step, we will first run an extensive measurement campaign to collect data from operational MBB networks in order to understand their performance and reliability. Second, we will derive machine-learning based models using the collected data in order to capture the network characteristics of MBB networks. Finally, we will extend our models to consider the performance of mobile video, more specifically video streaming, experienced by the end users. These models will be instrumental for different stakeholders from operators to application developers in improving the quality of their services and products. Therefore, MEMBRANE will have a significant impact on different sectors of industry while helping to improve the performance of their products leading to a better user experience for the end users.