University of Sarajevo | Faculty of Electrical Engineering | Coursware | Webmail

•

MONKS 2019 – Development of long-term programmable data plane architecture for 5G mobile network

BASIC INFORMATION

ABOUT THE PROJECT

Introduction

More than one billion devices worldwide are currently connected to fourth-generation (4G) mobile networks. However, it is already clear that, in the near future, the performance provided by these networks will be insufficient to meet the demand for connecting all devices to the Internet. As many as 100 billion connected devices are expected by 2025, including not only smartphones but also drones, autonomous vehicles, smart home devices and other equipment requiring higher data transfer rates, wider bandwidth and substantially lower latency than currently available technologies can provide. Developing a new generation of mobile networks—the fifth generation (5G)—has therefore become essential. These networks are attracting considerable attention worldwide, and intensive standardization efforts are underway.

Advanced 5G systems, envisioned as ultrafast networks providing instantaneous connectivity and underpinning the future Internet, represent not merely an evolution of the current generation of mobile networks but a revolution in information and communication technologies (ICT). ICT has long been one of the main drivers of the European economy, with sector revenues accounting for as much as 10% of total European GDP and an annual market value of nearly EUR 700 billion. In this context, 5G networks, currently among the most prominent ICT developments, are expected to provide highly reliable, ultrafast services everywhere and for everyone, thereby contributing both to global GDP growth and to improvements in quality of life worldwide. Achieving this requires a complete redesign of existing 4G interfaces, protocols and services, together with carefully developed algorithms and architectures for access and core networks.

Although 5G networks have already been deployed in several cities across Asia, America and Europe, available test results indicate that these networks have not fully achieved their intended performance targets. These include very low latency, which should be around 1 ms but has so far ranged from 8 to 12 ms, and data transfer rates in the Gbps range. Furthermore, deployed 5G networks have not met their targets for energy efficiency or architectural flexibility. Their architectures should be capable of accommodating current and future market requirements and diverse applications, including communication support for autonomous vehicles, smart homes and virtual reality. Consequently, the standardization process, expected to be completed by the end of next year, must place considerable emphasis on developing new concepts and architectures capable of delivering the planned 5G performance.

A particular challenge facing 5G networks concerns the radio segment. To achieve extremely high data transfer rates, 5G networks are intended to use the millimetre-wave band for the first time. This band has not yet been sufficiently investigated, either in terms of propagation or potential adverse effects on human health. During the earliest stages of testing proposed network architecture changes, and until medical investigations into the potential harmful effects of millimetre-wave use on human health have been completed, testing should ideally be conducted in controlled environments. In addition to the necessary architectural changes in the access and core segments of 5G networks, this creates a need for a suitable test environment capable of simulating realistic radio channel conditions.

Research topic

A review of the research and technical literature on 5G mobile networks shows that current development directions focus on adopting software-defined networking concepts. Under these concepts, base stations are simplified by moving most of their previous functions to the cloud, i.e. the network core. This process involves two steps: (1) softwarization and (2) virtualization. Softwarization means implementing network functions previously provided by specialized hardware and software as software applications running on conventional servers. Virtualization means running these software-based network functions in a virtual environment comprising server and network resources shared among many users. This improves server and network resource utilization and reduces base station costs. However, although this approach lowers the cost of the base station itself, it substantially reduces the flexibility of its functions, limiting potential changes needed to improve mobile networks as they evolve towards subsequent generations, such as 6G. Current research indicates that this limitation primarily affects base station functions related to data processing and transmission, i.e. data plane functions. By contrast, functions responsible for managing the network and its elements, i.e. control plane functions, were already successfully softwarized in the previous generation of mobile networks (4G) and are therefore no longer a research focus from the perspective of functional flexibility.

Accordingly, this project investigates the data plane architecture of 5G mobile networks, focusing on flexibility as a key factor in achieving 5G targets for maintenance costs, performance and energy consumption. Research on the data plane in softwarized networks has shown that a high degree of flexibility can only be achieved by ensuring long-term programmability. The research hypothesis is therefore that applying long-term programmability to the design of the 5G data plane architecture can bring networks closer to their targets for maintenance costs, performance and energy consumption, while paving the way for evolution towards 6G.

The number of research papers on the softwarization of 5G mobile networks—more than 1,000 papers over the past two years—indicates that the selected research topic is timely and important, with considerable potential to generate scientific publications. Since 5G networks represent the current step towards 6G, this research also offers opportunities for further development and expansion.

Project description

The project will propose a new approach to distributing functions between base stations and the 5G network core, retaining selected data processing and transmission functions at the base station as required. This would substantially improve the flexibility of the proposed architecture and could help achieve targets for reducing network maintenance and upgrade costs, improving performance and increasing energy efficiency. Maintenance and upgrade costs are expected to decrease significantly under the proposed architecture. To achieve this, a comparative analysis of existing research on the 5G data plane will first be conducted using multiple criteria to identify factors affecting maintenance costs, performance and energy consumption. Based on these factors, a new architectural model for the 5G data plane will be proposed, actively supporting 5G objectives through the long-term programmability of data plane elements.

The new architectural model will be experimentally verified by implementing a prototype of the 5G data plane in the Laboratory for Software-Defined Networks at the Faculty of Electrical Engineering, University of Sarajevo. Existing laboratory equipment, including servers, computer workstations and network switches, will be used to implement the network core. The base station will be implemented using a USRP X310 development platform, whose procurement is planned as part of the project.

Considering the challenges described above concerning the limitation of unnecessary and unwanted non-ionizing radiation, the project will develop a 5G radio channel simulator alongside its new approach to designing 5G network elements. The simulator will enable the proposed architectures to be tested in approximately realistic environments without potentially harmful effects on the surroundings. It will be based on a new geometry-based stochastic channel model offering a more unified approach than existing models and enabling the description and simulation of realistic propagation environments in densely populated and urban areas.

In view of the above, and the growing emphasis on reducing energy consumption and radio emissions, particularly in insufficiently explored parts of the spectrum, the proposed project is expected to deliver benefits across several areas of society.

Project objective

The project aims to reduce maintenance costs, improve performance and increase energy efficiency in fifth-generation (5G) mobile networks by implementing data plane components with long-term programmability.

Expected research outcomes

The expected project outcomes are divided into scientific and practical outcomes. The scientific outcomes can be summarized as follows:

  • R1: A new original metric enabling the identification of architectural factors affecting maintenance costs, performance and energy consumption in 5G mobile networks through a comparative analysis of 5G data plane research using multiple criteria.
  • R2: A new architectural model for the 5G data plane that, through long-term programmability, actively supports lower maintenance costs, improved performance across the entire 5G mobile network and more efficient energy use.
  • R3: A new geometry-based stochastic channel model offering a more unified approach than existing channel models, without requiring the questionably realistic assumption that scatterers are uniformly distributed within specific two-dimensional or three-dimensional regions.
  • R4: A new simulation model of the 5G mobile communication channel, based on the new geometry-based stochastic channel model.

The expected practical outcomes are:

  • R5: A prototype of the new 5G data plane architecture, based on communication components with long-term programmability, such as switches and base stations.
  • R6: A new protocol for configuring the new 5G data plane architecture.
  • R7: A 5G mobile network channel simulator enabling experimental evaluation of the characteristics of the prototype of the new 5G data plane architecture.

The significance of the expected outcomes is as follows:

  • Developing the new original metric (R1) will establish a basis for optimizing the design of the 5G data plane to address all aspects of the project objective: reducing maintenance costs, improving performance and increasing energy efficiency. This will lead to the development of a new architectural model for the 5G data plane (R2). The prototype of the new architecture (R5) and its configuration protocol (R6) will enable experimental verification of the project results and establish a laboratory environment for future research on 5G mobile networks.
  • The new communication channel model (R3) will make it possible to describe a substantially wider range of realistic environments than those covered in the existing literature. In addition, the channel simulation model (R4) will enable the simulation of realistic propagation environments of particular interest in densely populated and urban areas. By optimizing base station parameters in 5G mobile networks, these simulations will support efforts to reduce unnecessary and unwanted non-ionizing radiation and improve energy efficiency. The channel simulator (R7) will provide a practical laboratory tool for conducting these simulations, both within this project and in future research on 5G mobile networks.
640px-NATO_OTAN_landscape_logo.svg-2627358850
previous arrowprevious arrow
next arrownext arrow