Virtual network embedding in multi-domain environments with energy efficiency concepts

Araújo, Samuel Moreira Abreu; Souza, Fernanda Sumika Hojo de; Mateus, Geraldo Robson

doi:10.1109/icoin.2018.8343111

Cited by 6 publications

(10 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The problem of embedding this graph into the physical network towards some optimization objective (e.g., minimization of the deployment cost) is referred to as Virtual Network Embedding (VNE) and has received considerable attention in literature [1]. In particular, the general VNE problem was proved to be NP-hard [2]; hence, many heuristics have been proposed to efficiently solve it [3]. VNE is a wellstudied problem when a single network operator is considered, but, to the extent of our knowledge, the case where the underlying infrastructure is composed of several independent ISPs' networks has only been scarcely explored so far.…”

Section: Andreoletti and T Velichkova Contributed Equally To This Wmentioning

confidence: 99%

A Privacy-Preserving Reinforcement Learning Algorithm for Multi-Domain Virtual Network Embedding

Andreoletti

Velichkova

Verticale

et al. 2020

IEEE Trans. Netw. Serv. Manage.

View full text Add to dashboard Cite

The problem of optimally deploying a virtual network onto a substrate physical network is referred to as Virtual Network Embedding (VNE). In general, this embedding is requested by a customer to an Internet Service Provider (ISP), which performs the VNE over its physical telecom network. In several situations, the physical substrate infrastructure is composed of multiple independent ISPs. In this scenario, ISPs are concerned about exposing to a third-party entity (e.g., the customer) sensitive infrastructural details that are needed to perform an effective embedding. Following a common privacypreserving approach, known as Limited Information Disclosure (LID), the embedding may be performed by the customer based on a limited and abstracted view of the multi-domain infrastructure that ISPs accept to expose. With this approach, embedding is sub-optimal (e.g., embedding cost is not minimized) in comparison with the case where all information is available, i.e., Full Information Disclosure (FID). In this work, we propose a Reinforcement-Learning-based algorithm able to process data that the customer and ISPs cipher under the Shamir Secret Sharing (SSS) scheme. This approach guarantees total privacy to both the customer and the ISPs (e.g., details about a virtual function are only revealed to the ISP in charge of hosting it) and achieves comparable embedding cost of an existing FID heuristic, as observed from extensive simulations. The main drawback of our algorithm is the high overhead of data that ISPs and the customer need to exchange with each other to execute it. Hence, we also explore the trade-off between embedding cost and data overhead resulting from the reduction of operations done by the RL. In general, intermediate embedding costs between the FID and LID heuristics can be obtained at a significant reduction of data overhead, while not sacrificing any privacy guarantees.

show abstract

Section: Andreoletti and T Velichkova Contributed Equally To This Wmentioning

confidence: 99%

A Privacy-Preserving Reinforcement Learning Algorithm for Multi-Domain Virtual Network Embedding

Andreoletti

Velichkova

Verticale

et al. 2020

IEEE Trans. Netw. Serv. Manage.

View full text Add to dashboard Cite

show abstract

“…Zhang [12] presented a new node ranking scheme and differentiated pricing strategy to evaluate the node and link weights during the VNE process, which achieved a better trade-off between the energy consumption and network load balance. A hybrid algorithm was studied in [13]. This algorithm combined the greedy randomized adaptive search procedure and the reduced variable neighborhood search method to obtain the better energy efficiency in VNE.…”

Section: Related Workmentioning

confidence: 99%

Energy Optimized Virtual Network Embedding With Location Constraint in the Enterprise Network

Cong

Shuang

2020

IEEE Access

View full text Add to dashboard Cite

To satisfy enterprise demands of analyzing and dealing with the large scale of data with lower costs, an effective method is to integrate the servers and computers and use virtualization technology to construct an enterprise network. Prior studies on network virtualization have mainly been executed in the cloud; however, these studies may not be appropriate for enterprise networks for two reasons: i) the goal of most of them is to generate more revenues for cloud providers, but focus less on saving costs; ii) the physical machines are relatively concentrated in the cloud platform but dispersed over different geographic locations in enterprise networks. In this paper, we solve the problem of energy-optimized virtual network embedding with location constraints (EO-VNE). First, the node and link capabilities in enterprise networks are defined in the form of complex number theory, unifying computers and virtual requests. Second, the normalized method of computing and storage capabilities are proposed to identify the node capability. Third, an energy model of the enterprise network is built, and using this model, EO-VNE is shown to be NP-complete. Finally, an energy-optimized virtual network embedding with a location constraint algorithm (EOLC) is proposed to minimize the energy consumption under the constraint of node position. The experiments show that EOLC consumes less energy compared with the algorithm of energy-aware virtual network embedding with dynamic demands (EAD). It also has better performance than the location constraint algorithm based on bisection (GLC).

show abstract

“…Essa indagação induz à formulação da seguinte questão de pesquisa: Como identificar quando se deve aplicar um tratamento heurístico ou exato ao VNE? Nesse aspecto, juntamente aos preceitos de [Shi et al 2015] e [Araújo et al 2018], foi considerado que uma requisição sempre deve ser atendida com um baixo tempo de processamento.…”

Section: Processamento Heurístico E Exatounclassified

“…Neste caso, a tomada de decisão sobre a aceitação ou rejeição da requisição deve ser realizada de maneira otimizada e rápida, de acordo com a disponibilidade de recursos residuais do SN no momento do mapeamento. Outra característica do problema é a não aplicação de reconfiguração, diante disso, cada requisição perdura com o mesmo mapeamento sobre o SN até ter seu tempo de vida expirado (mapeamento estático), aspectos aplicados em [Chowdhury et al 2009, Beikta et al 2017, Araújo et al 2018.…”

Section: Introductionunclassified

“…Seja no problema VNE, ou no NFV, a alocação de recursos em nuvem requer uma resposta de processamento rápida para se manter níveis aceitáveis de Qualidade de Serviço (QoS) [Shi et al 2015, Araújo et al 2018. Ressalta-se que o tempo de processamento máximo que uma requisição de rede pode aguardar é um parâmetro de QoS atribuído ao usuário final, i.e., cada aplicação pode aguardar um tempo máximo diferente para ser processada.…”

Section: Introductionunclassified

See 1 more Smart Citation

Alocação de Recursos para Redes Virtuais com Seleção de Método de Resolução via Aprendizado de Máquina

Araújo¹,

Souza²,

Mateus³

2020

Anais XXXVIII Simpósio Brasileiro De Redes De Computadores E Sistemas Distribuídos (SBRC 2020)

View full text Add to dashboard Cite

A alocação de recursos é um problema NP-difícil, presente no Mapeamento de Redes Virtuais (VNE) e na Virtualização de Funções de Rede (NFV). Nesse contexto, um ponto pouco explorado na literatura é relacionado a quando se deve aplicar um método de resolução heurístico ou exato. Tanto para o VNE, quanto para o NFV, normalmente a literatura indica um tratamento heurístico devido à complexidade do problema e à elevada dimensão dos dados. Entretanto, em experimentos preliminares, percebe-se que em alguns momentos um tratamento exato pode ser aplicado em um tempo praticável. Foi notado que esses eventos não se tratam de casualidades, mas decorrem de várias condições relacionadas ao Substrato de Rede (SN) residual e às demandas de rede requisitadas. A abordagem proposta neste artigo é baseada na aplicação de técnicas de Aprendizado de Máquina (AM) para alocação de recursos em rede, e tem como objetivo prever em quais situações se pode preterir uma abordagem heurística em prol de uma exata. Para tal, foi adotado o VNE e duas abordagens da literatura. As simulações realizadas mostram que o tratamento com AM melhora a taxa de aceitação e a receita de mapeamento em relação à heurística, e reduz o tempo de processamento em relação à abordagem exata.

show abstract

Virtual network embedding in multi-domain environments with energy efficiency concepts

Cited by 6 publications

References 17 publications

A Privacy-Preserving Reinforcement Learning Algorithm for Multi-Domain Virtual Network Embedding

A Privacy-Preserving Reinforcement Learning Algorithm for Multi-Domain Virtual Network Embedding

Energy Optimized Virtual Network Embedding With Location Constraint in the Enterprise Network

Alocação de Recursos para Redes Virtuais com Seleção de Método de Resolução via Aprendizado de Máquina

Contact Info

Product

Resources

About