Network virtualization in next generation cellular networks
The complexity of operation and management of emerging cellular networks significantly increases, as they evolve to correspond to increasing QoS needs, data rates and diversity of offered services. Thus critical challenges appear regarding their performance. At the same time, network sustainability...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis doctoral |
| Fecha de publicación: | 2019 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/166962 |
| Acceso en línea: | https://hdl.handle.net/2117/166962 https://dx.doi.org/10.5821/dissertation-2117-166962 |
| Access Level: | acceso abierto |
| Palabra clave: | Comunicacions mòbils, Sistemes de Xarxes neuronals (Informàtica) Àrees temàtiques de la UPC::Enginyeria de la telecomunicació |
| Sumario: | The complexity of operation and management of emerging cellular networks significantly increases, as they evolve to correspond to increasing QoS needs, data rates and diversity of offered services. Thus critical challenges appear regarding their performance. At the same time, network sustainability pushes toward the utilization of haring Radio Access Network (RAN) infrastructure between Mobile Network Operators (MNOs). This requires advanced network management techniques which have to be developed based on characteristics of these networks and traffic demands. Therefore it is necessary to provide solutions enabling the creation of logically isolated network partitions over shared physical network infrastructure. Multiple heterogeneous virtual networks should simultaneously coexist and support resource aggregation so as to appear as a single resource to serve different traffic types on demand. Hence in this thesis, we study RAN virtualization and slicing solutions destined to tackle these challenges. In the first part, we present our approach to map virtual network elements onto radio resources of the substrate physical network, in a dense multi-tier LTE-A scenario owned by a MNO. We propose a virtualization solution at BS level, where baseband modules of distributed BSs, interconnected via logical point-to-point X2 interface, cooperate to reallocate radio resources on a traffic need basis. Our proposal enhances system performance by achieving 53% throughput gain compared with benchmark schemes without substantial signaling overhead. In the second part of the thesis, we concentrate on facilitating resource provisioning between multiple Virtual MNOs (MVNOs), by integrating the capacity broker in the 3GPP network management architecture with minimum set of enhancements. A MNO owns the network and provides RAN access on demand to several MVNOs. Furthermore we propose an algorithm for on-demand resource allocation considering two types of traffic. Our proposal achieves 50% more admitted requests without Service Level Agreement (SLA) violation compared with benchmark schemes. In the third part, we devise and study a solution for BS agnostic network slicing leveraging BS virtualization in a multi-tenant scenario. This scenario is composed of different traffic types (e.g., tight latency requirements and high data rate demands) along with BSs characterized by different access and transport capabilities (i.e., Remote Radio Heads, RRHs, Small Cells, SCs and future 5G NodeBs, gNBs with various functional splits having ideal and non-ideal transport network). Our solution achieves 67% average spectrum usage gain and 16.6% Baseband Unit processing load reduction compared with baseline scenarios. Finally, we conclude the thesis by providing insightful research challenges for future works. |
|---|