Back to search

University of Thessaly (UTH)

Design and experimental study of heterogeneous disaggregated 5G mobile wireless networks

Abstract

dc:description

The fifth generation of mobile networking (5G) is fostering several advancements in the access, edge and core network, promising to offer significantly higher network capacity with lower latency over the network, allowing a variety of applications and critical services to thrive around this ecosystem. Extensive softwarization of the network adds up to the flexibility of management and eases future upgrades. Softwarization expands even to the Radio Access Network (RAN) through the emerging concept of Cloud-RAN. In such setups, part of the base station stack is running as a software unit that can be instantiated in the Cloud, controlling different radio access modules for providing network access to the network. Relevant literature suggests different points for splitting the base station stack, with each point bringing different benefits for the operator. As we move down to the network stack, splits in the lower physical layer pose stringent requirements (very low latency, very high capacity) for the transport of signaling and data to the cloud-located unit (fronthaul network). Splits taking place at higher layers may prove to be more beneficial for the operator, as they can be served with existing network infrastructure (e.g. Ethernet links), and also allow the integration of heterogeneous technologies to the cell. Integration with existing technologies at different levels (access and transport network) is crucial for the continuous evolution of the platform and compatibility with legacy protocols. In this thesis, we design, study and introduce novel functionality in disaggregated base station setups, towards enabling the operation of heterogeneous technologies in the user access. The fundamental questions that we try to answer are the following: 1) What is the best point to disaggregate the base station stack? 2) How can heterogeneous technologies be integrated in the network cell? 3) Given such a disaggregated infrastructure, how can we serve end users with low latency? 4) Since such infrastructure is highly softwarized, how can we efficiently orchestrate its operation? Initially we begin with the cellular stack, and study and design its disaggregation at different points. Depending on the point that the stack is dissaggregated, different requirements are posed on the transport network, from the radio unit to the cloud. As we also target ntegration with existing technologies for the transport network (e.g. Ethernet/WiFi), we design and develop splits for the stack that can be accommodated using off-the-shelf equipment. We use a packetized transport network (base station fronthaul), in order not to rely on expensive hardware or newer protocols (e.g. the Common Public Radio Interface - CPRI). We develop and implement our splits for the stack in the open source framework OpenAirInterface (OAI), and experimentally evaluate different splits in a testbed setup. Following this, we delve into the higher OSI layer 2 split for the cellular stack and design and study the integration of different technologies to the cell. The higher layer splits that are currently standardized for the 5G New Radio (NR) protocol, assume the disaggregation of the stack at the same point, between the Packet Data Convergence Protocol (PDCP) and the Radio Link Control (RLC) layers. This disaggregation creates two different units, the Centralized Unit (CU) running at the Cloud, and the Distributed Unit (DU) running the lower layer 2 and Layer 1 protocols. Nevertheless, the PDCP layer has been used in the past for integrating different technologies in the cell, such as for example the LTE-WiFi Aggregation Adaptation Protocol (LWAAP). This creates the circumstances for the integration of heterogeneous technologies in the cell as DUs. Based on our prior study, we design and integrate WiFi as a radio access technology to the base station cell. Using a load-balancing controller, we can manage in a per packet basis which DU can serve each network client. We evaluate our contributions by extending the OAI prototype implementation and experiment in real settings with the prototype. Following this, we provide performance metrics when tuning different parameters (e.g. distance between CU/DU, transport protocol used, etc.). Nevertheless, such infrastructures tend to offer ultra-dense network deployment, with multiple technologies being collocated within the same spectrum. Therefore, efficient coordination needs to take place in order to ensure the proper operation of all the accommodated technologies. To this aim, we design and study a spectrum coordination scheme that can be applied in such disaggregated heterogeneous base station architectures. We design signaling that carries radio configuration parameters and statistics, in order to conclude on the spectrum usage of a specific area. Based on the collected measurements, we apply an algorithm for efficiently placing all the technologies within the same frequency band. We evaluate our design by further extending the implementation and experimenting in a real testbed environment with densely deployed wireless networks. Our results illustrate higher performance of the under-study network. As our next step, we design and study new features in such multi-technology disaggregated networks, towards providing low-latency service access for users connected to the cell. Multi-access Edge Computing (MEC) has been proposed as a means to drastically minimize the service access latency for clients connected to the network, by placing the services closer to the network edge. Existing literature suggests different placements for the MEC services, which at the best case is collocated with the CU of the network, intercepting backhaul traffic. Given the disaggregated nature of the cell, we design and study how the placement of services on the fronthaul network, collocated with/close to the DUs of the network, can reduce the service access latency. Our testbed experiments denote that in terms of latency, even legacy technologies (e.g. 4G LTE) can be suitable for several of the proposed 5G applications. Finally, as this network is highly softwarized, we design and study the orchestration of the software as Virtual Network Functions (VNFs). Network Functions Virtualization Management and Orchestration (NFV-MANO) provides a standardized approach on the management and effortless deployment of (virtual) services. Although NFV-MANO initially focused on the deployment of services over datacenters, the introduction of fully softwarized network architectures even for the wireless part creates fertile ground for the re-conception of the manner through which the underlying hardware is managed. In this concept, we design extensions to a Virtual Infrastructure Manager (VIM) in order to handle virtualized wireless network interfaces, hosted on the generic networking nodes of the testbed. We design the extensions in a manner that allows their transparent introduction to the existing operation of the platform, thus allowing portability of network services and network functions to other instances as well. We implement and apply our design in a testbed environment, and subsequently benchmark the framework in terms of performance and evaluate it by using the disaggregated multi-technology base stations that we have previously developed.

Degree

thesis:*
Grantor dc:publisher
University of Thessaly (UTH)
Year dc:date
2020

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Makris, Nikolaos
  • Μακρής, Νικόλαος

Subjects

dc:subject × 15

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/47054

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Makris, Nikolaos; Μακρής, Νικόλαος. Design and experimental study of heterogeneous disaggregated 5G mobile wireless networks. University of Thessaly (UTH), 2020. http://hdl.handle.net/10442/hedi/47054