Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml Jakob Voss, based on art designer at PLoS, modified by Wikipedia users Nina and Beao Closed Access logo, derived from PLoS Open Access logo. This version with transparent background. http://commons.wikimedia.org/wiki/File:Closed_Access_logo_transparent.svg Jakob Voss, based on art designer at PLoS, modified by Wikipedia users Nina and Beao Estudo Geralarrow_drop_down
image/svg+xml Jakob Voss, based on art designer at PLoS, modified by Wikipedia users Nina and Beao Closed Access logo, derived from PLoS Open Access logo. This version with transparent background. http://commons.wikimedia.org/wiki/File:Closed_Access_logo_transparent.svg Jakob Voss, based on art designer at PLoS, modified by Wikipedia users Nina and Beao
Estudo Geral
Doctoral thesis . 2025
Data sources: Estudo Geral
ResearchGate Data
Thesis . 2025
Data sources: Datacite
versions View all 2 versions
addClaim

Optimising Resource Allocation in Fog Computing

Authors: Godinho, Noé Paulo Lopes;

Optimising Resource Allocation in Fog Computing

Abstract

Devido ao grande aumento de dispositivos da Internet das Coisas nos últimos anos, a núvem tem-se revelado inadequada para lidar com o elevado número de pedidos gerados por esses dispositivos. Para resolver este problema, foi proposta a computação em névoa, que fornece serviços de computação mais próximos de forma distribuída, atuando como uma camada intermédia entre os dispositivos da Internet das Coisas e a núvem. No entanto, estes dispositivos possuem baterias com capacidade pequena e baixo poder computacional. Por conseguinte, estratégias para distribuir os serviços solicitados, agrupados num conjunto de aplicações, são de particular interesse. Estas aplicações podem ter prazos curtos, que precisam de ser cumpridos, e outros requisitos de interesse que precisam de ser assegurados, geralmente na forma de Qualidade de Serviço. Assim, existe a necessidade de resolver problemas de Colocação de Serviços com diferentes métricas e restrições.Devido à heterogeneidade dos dispositivos em ambientes em névoa, a Virtualização de Redes surgiu naturalmente. Neste paradigma, o objetivo é virtualizar recursos para permitir isolamento e facilitar a conectividade entre dispositivos físicos, utilizando software como uma camada de abstração. Um passo além é o Fatiamento de Redes, onde a própria rede é virtualizada para proporcionar segurança, uso eficiente de recursos e facilitar a configuração dos dispositivos. Ambos os paradigmas necessitam de alocar um conjunto de Redes Virtuais de forma eficiente, de acordo com métricas de interesse.A Reconfiguração e Migração é um problema comum existente em ambientes de computação em névoa. Não só os utilizadores podem mover-se constantemente, tornando necessário garantir uma conectividade contínua, mas também os dispositivos na rede podem ficar indisponíveis, devido ao esgotamento da bateria ou ao uso excessivo. Neste cenário, é necessário garantir uma realocação de caminhos e recursos para assegurar uma Qualidade de Serviço adequada.Nesta tese, o objetivo é abordar a seguinte questão: como modelar e resolver estes difíceis problemas de alocação de recursos? Para cada um dos problemas descritos anteriormente, é desenvolvida uma formulação de Programação Linear Inteira Mista, uma heurística com múltiplas métricas a otimizar e várias restrições associadas. Dado que os problemas estão interligados, é também proposta uma framework que junta os três tipos de problemas. Para avaliar cada uma destas abordagens, é realizada uma análise experimental e uma simulação. As heurísticas são comparadas com as abordagens do estado da arte. As soluções obtidas são próximas do valor ótimo e melhores do que os algoritmos do estado da arte.

Due to the large increase of Internet of Things devices in the last years, the cloud has been proved unsuitable to deal with the high demand of requests generated by them. To deal with this, fog computing was proposed to provide closer computing services in a distributed manner, acting as a middle layer between Internet of Things devices and the cloud. However, these devices have small battery capacity and low computational power. Therefore, strategies to distribute requested services, bundled in a set of applications, are of particular interest. These applications may have deadlines that are short, which have to be met, and other requirements of interest that need to be ensured, usually in the form of Quality of Service. Thus, there is a need to solve Service Placement problems by taking into account different metrics and constraints.Due to the heterogeneity of devices in fog environments, Network Virtualisation emerged naturally. In this paradigm, the goal is to virtualise resources in order to allow isolation and easier connectivity between physical devices by software as an abstraction layer. A step further is Network Slicing, where the network itself is virtualised to provide security, resource usage and ease the configuration of devices. In both paradigms, there is a need to efficiently allocate a set of Virtual Networks efficiently, according to metrics of interest.Reconfiguration and Migration is a common problem to solve in fog computing environments. Not only users may move constantly, making it necessary to ensure constant connectivity, but devices in the network may be unavailable, due to battery depletion or overusage. In this scenario, there is a need to ensure a relocation of paths and resources in order to ensure suitable Quality of Service.In this thesis, the goal is to tackle the following question: how to model and solve these hard resource allocation problems? For each of the previously described problems, a Mixed Integer Linear Programming formulation and heuristic are developed with multiple metrics to optimise and several associated constraints. Since the problems are interconnected, a framework joining all three types of problems is proposed as well. To evaluate each of these approaches, simulation and an in-depth experimental analysis is performed. The heuristics are compared to state-of-the-art approaches. The computational results show that our heuristics are close to the optimal value and better than known state-of-the-art algorithms.

Outro - OREOS: Orchestration and Resource optimization for rEliable and lOw-latency serviceS - https://www.uc.pt/administracao/dpa/investigacao/proj_cof/oreos

Outro - SNOB - 5G: Scalable Network Backhauling for 5G - https://www.uc.pt/administracao/dpa/investigacao/proj_cof/SNOB_5G_045929

Outro - 5G - Componentes e serviços para redes 5G - https://www.uc.pt/administracao/dpa/investigacao/proj_cof/5g

Tese de Doutoramento em Engenharia Informática apresentada à Faculdade de Ciências e Tecnologia

FCT

Country
Portugal
Related Organizations
Keywords

Qualidade de Serviço, Multi-objective Optimisation, Ciências da engenharia e tecnologias::Engenharia electrotécnica, electrónica e informática, Alocação de Recursos, Otimização Multi-objetivo, Internet of Things, Quality of Service, Internet das Coisas, Computação em Névoa, Fog Computing, Resource Allocation

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
0
Average
Average
Average
Upload OA version
Are you the author of this publication? Upload your Open Access version to Zenodo!
It’s fast and easy, just two clicks!