Go to the content. | Move to the navigation | Go to the site search | Go to the menu | Contacts | Accessibility

| Create Account

Dlamini, Thembelihle (2019) Core Network Management Procedures for Self-Organized and Sustainable 5G Cellular Networks. [Ph.D. thesis]

Full text disponibile come:

[img]
Preview
PDF Document - Accepted Version
Available under License Creative Commons Attribution Non-commercial.

1440Kb

Abstract (italian or english)

This thesis investigates resource management procedures, within the Multi-access Edge Computi ng (MEC) paradigm, to obtain energy savings and guarantee Quality of Service(QoS) in Mobile Networks (MNs). Here, we enable energy savings within green-aware network apparatuses (i.e., communication and computing facilities) through the application of learning and control techniques, together with energy management procedures (BS sleep mode, VM soft-scaling, tuning of transmission drivers). In this study, we consider the MEC deployment scenarios suggested by ETSI and mobile operators for our system models.

Firstly, we investigate energy-saving strategies within a remote site fully powered by only green/renewable energy (solar and wind). Here, we consider a single Base Station (BS) co-located with the MEC server, i.e., the BS is empowered with computing capabilities. To address the energy consumption problem within the remote site, we propose an online algorithm for edge network management. The algorithm make use of a Long Short-Term Memory (LSTM) neural network for estimating the short-term future traffic load and harvested energy, and control theory, specifically the Limited Lookahead Control (LLC) principles, for foresighted optimization. It also make use of energy management procedures, i.e., BS sleep modes and Virtual Machine (VM) soft-scaling (the reduction of computing resources per time instance). To obtain the energy savings and guarantee QoS, per time instance, the algorithm considers the future BS loads, onsite green energy available and then provisions edge network resources based on the learned information.

Secondly, we study the energy consumption problem within an environment where BSs are densely-deployed, i.e., similar to an urban or semi-urban scenario. This work extend the energy consumption problem from a single BS case to multiple BSs. Here, each BS is powered by hybrid energy supplies (solar and power grid) and also empowered with computation capabilities (each BS is co-located with a MEC server). Towards edge system management, we propose a controller-based network architecture for managing energy harvesting (EH) BSs empowered with computation capabilities where on/off switching strategies allow BSs and VMs to be dynamically switched on/off, depending on the traffic load and the harvested energy forecast, over a given look-ahead prediction horizon. To solve the energy consumption minimization problem in a distributed manner, the controller partitions the BSs into clusters based on their location; then, for each cluster, it minimizes a cost function capturing the individual communication site energy consumption and the users’ QoS. To manage the communication sites, the controller performs online supervisory control by forecasting the traffic load and the harvested energy using a LSTM neural network, which is utilized within a LLC policy to obtain the system control actions that yield the desired trade-off between energy consumption and QoS.

Finally, we investigate the energy consumption problem within a virtualized MEC server placed in proximity to a group of BSs. To address this challenge, we consider a computing-plus-communication energy model, within the MEC paradigm, where we focus on the communication-related energy cost in addition to the energy drained due to computing processes. Towards server management, an online algorithm based on traffic engineering and MEC Location Service is proposed. To obtain the energy savings and QoS guarantee, we jointly launch an optimal number of VMs for computing and transmission drivers coupled with the location-aware traffic routing for real-time data transfers. In order to efficiently provisioned edge system resources, we forecast the server workloads and harvested energy by using a LSTM neural network and the output is then used within the LLC-based algorithm.

Our numerical results, obtained through trace-driven simulations, show that the proposed optimization strategies (algorithms) leads to a considerable reduction in the energy consumed by the edge computing and communication facilities, promoting energy self-sustainability within the MN through the use of green energy.


Statistiche Download
EPrint type:Ph.D. thesis
Tutor:Rossi, Michele
Ph.D. course:Ciclo 32 > Corsi 32 > INGEGNERIA DELL'INFORMAZIONE > SCIENZA E TECNOLOGIA DELL'INFORMAZIONE
Data di deposito della tesi:30 November 2019
Anno di Pubblicazione:30 November 2019
Key Words:Multi-access edge computing, energy harvesting, soft-scaling, adaptive control
Settori scientifico-disciplinari MIUR:Area 09 - Ingegneria industriale e dell'informazione > ING-INF/03 Telecomunicazioni
Struttura di riferimento:Dipartimenti > Dipartimento di Ingegneria dell'Informazione
Codice ID:12215
Depositato il:22 Jan 2021 12:38
Simple Metadata
Full Metadata
EndNote Format

Download statistics

Solo per lo Staff dell Archivio: Modifica questo record