DROI: Energy-efficient virtual network embedding algorithm based on dynamic regions of interest

He, Mengyang; Zhuang, Lei; Tian, Shuaikui; Wang, Guoqing; Zhang, Kunli

doi:10.1016/j.comnet.2019.106952

Cited by 13 publications

(3 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, this scheme cannot be applied to large-scale networks due to the limitations of the heuristics. In response to the lack of consideration of network topology information in energy-efficient VNE, field theory-based spectral clustering is applied to extract network features [28]. The modified method builds dynamic regions of interest based on the topological information of the network and then updates the mappable regions in real time, avoiding the local optimum problem.…”

Section: Energy-efficient Virtual Network Embeddingmentioning

confidence: 99%

Energy-Efficient Virtual Network Embedding: A Deep Reinforcement Learning Approach Based on Graph Convolutional Networks

Zhang,

Wang,

Luo

et al. 2024

Electronics

View full text Add to dashboard Cite

Network virtualization (NV) technology is the cornerstone of modern network architectures, offering significant advantages in resource utilization, flexibility, security, and streamlined management. By enabling the deployment of multiple virtual network requests (VNRs) within a single base network through virtual network embedding (VNE), NV technology can substantially reduce the operational costs and energy consumption. However, the existing algorithms for energy-efficient VNE have limitations, including manual tuning for heuristic routing policies, inefficient feature extraction in traditional intelligent algorithms, and a lack of consideration of periodic traffic fluctuations. To address these limitations, this paper introduces a novel approach that leverages deep reinforcement learning (DRL) to enhance the efficiency of traditional methods. We employ graph convolutional networks (GCNs) for feature extraction, capturing the nuances of network graph structures, and integrate periodic traffic fluctuations as a key constraint in our model. This allows for the predictive embedding of VNRs that is both energy-efficient and responsive to dynamic network conditions. Our research aims to develop an energy-efficient VNE algorithm that dynamically adapts to network traffic patterns, thereby optimizing resource allocation and reducing energy consumption. Extensive simulation experiments demonstrate that our proposed algorithm achieves an average reduction of 22.4% in energy consumption and 41.0% in active substrate nodes, along with a 23.4% improvement in the acceptance rate compared to other algorithms.

show abstract

Section: Energy-efficient Virtual Network Embeddingmentioning

confidence: 99%

Energy-Efficient Virtual Network Embedding: A Deep Reinforcement Learning Approach Based on Graph Convolutional Networks

Zhang,

Wang,

Luo

et al. 2024

Electronics

View full text Add to dashboard Cite

show abstract

“…They modeled the VNE problem as a two-stage mapping problem and introduced a VNE algorithm called OPaCoVNE to solve the resource management problem while considering the end-to-end delay as the embedded constraint. However, due to the under-exploitation of physical resources in the substrate network, its resource utilization efficiency is low..He et al [17] used spectral clustering based on field theory to extract substrate network features and manage physical resources. Then, they developed dynamic regions of interest to find embedding areas with energy-saving potential for virtual networks.…”

Section: Related Workmentioning

confidence: 99%

An edge-based approach for virtual network embedding based on the graph edit distance

Zhuang

et al. 2023

PLoS ONE

View full text Add to dashboard Cite

Virtualization and resource isolation techniques have enabled the efficient sharing of networked resources. How to control network resource allocation accurately and flexibly has gradually become a research hotspot due to the growth in user demands. Therefore, this paper presents a new edge-based virtual network embedding approach to studying this problem that employs a graph edit distance method to accurately control resource usage. In particular, to manage network resources efficiently, we restrict the use conditions of network resources and restrict the structure based on common substructure isomorphism and an improved spider monkey optimization algorithm is employed to prune redundant information from the substrate network. Experimental results showed that the proposed method achieves better performance than existing algorithms in terms of resource management capacity, including energy savings and the revenue-cost ratio.

show abstract

“…Chen et al [22] established a feedback system that can identify the activity subset of physical resources, in which virtual networks are mapped to active areas and energy consumption is reduced by increasing the number of nodes and links in hibernation. He et al [23] used spectral clustering to extract physical network features, so as to distinguish the dynamic regions of interest and find embedding regions with energy-saving potential of virtual networks.…”

Section: Dynamic Vne Algorithmmentioning

confidence: 99%

DVNE-DRL: Dynamic virtual network embedding algorithm based on deep reinforcement learning

Xiao¹

2023

Preprint

View full text Add to dashboard Cite

Virtual Network Embedding (VNE), as the key challenge of network resource management technology, lies in the contradiction between online embedding decision and pursuing long-term average revenue goals. Most of the previous work ignored the dynamics in Virtual Network (VN) modeling, or could not automatically detect the complex and time-varying network state to provide a reasonable network embedding scheme. In view of this, we model a network embedding framework where the topology and resource allocation change dynamically with the number of network users and workload, and then introduce a deep reinforcement learning method to solve the VNE problem. Further, a dynamic virtual network embedding algorithm based on Deep Reinforcement Learning (DRL), named DVNE-DRL, is proposed. In DVNE-DRL, VNE is modeled as Markov Decision Process (MDP), and Reinforcement Learning Agent (RLA) is used to periodically detect the state of network environment and provide a more reasonable network embedding scheme in real time. Different from the previous methods based on DRL algorithm to solve VNE problem, this paper improves the method of feature extraction and matrix optimization, and considers the characteristics of virtual network and physical network together to alleviate the problem of redundancy and slow convergence. The simulation results show that compared with the existing advanced algorithms, the acceptance rate and average revenue of DVNE-DRL are increased by about 25% and 35%, respectively. K E Y W O R D S VNE, DRL, embedding framework, DRL-DVNE，MDP，feature extraction

show abstract

DROI: Energy-efficient virtual network embedding algorithm based on dynamic regions of interest

Cited by 13 publications

References 26 publications

Energy-Efficient Virtual Network Embedding: A Deep Reinforcement Learning Approach Based on Graph Convolutional Networks

Energy-Efficient Virtual Network Embedding: A Deep Reinforcement Learning Approach Based on Graph Convolutional Networks

An edge-based approach for virtual network embedding based on the graph edit distance

DVNE-DRL: Dynamic virtual network embedding algorithm based on deep reinforcement learning

Contact Info

Product

Resources

About