Towards intelligent virtual resource allocation in UAVs-assisted 5G networks
Introduction
Both academia and industry have reached the consensus that the 5G networks [1], [2], [3], [4] are designed for providing massive network elements connection and fulfilling explosive data growth. Meanwhile, more novel network services and applications [5], having high data rate and low latency, emerge in 5G era. Therefore, it is necessary to manage and schedule the physical network resources efficiently. Considering the heterogeneity nature of dedicated hardware, it is a great burden [6] for telecommunication service providers (TSPs) to invest and upgrade these hardware continuously. Therefore, virtualization technologies (e.g. network virtualization, NV [1], network function virtualization, NFV [4]) are seen as the most potential approach towards the upgrading tendency of 5G. By conducting virtualization, underlying physical resources (e.g. radio, computing, storage, bandwidth) can be abstracted into virtual resources easily. TSPs can manage, schedule and allocate their owned virtualized physical resources easily.
In order to implement 5G network services and load virtualized network elements to their full capacities, TSPs need to be equipped with abundant algorithms, having the function of allocating virtual resources efficiently. In the literature, it attracts significant research attention. In both academia and industry, the resource allocation problem can be called as virtual network embedding (VNE) [7]. There exist abundant technical publications in the literature [7], [8], [9], [10]. These publications enables to help following researchers to start the VNE research quickly. However, existing technical publications [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28] allocate the virtual resources in an inflexible way. For instance, once one network service is allocated and implemented, no intelligent scheme exists in order to adjust the allocation results. In addition, all nodes of virtual network services are assumed to be static throughout their lifespan. None considered the embedding and allocation in case certain node moves. Though a few publications (e.g. [27], [28]) discussed the node mobility, none mentioned which concrete network scenario [27] can be applied. However, ‘the last one mile’ in resource allocation aspect is very important. That is to say, the node mobility cannot be ignored in real networking research. In addition, existing publications (e.g. [28]) did not evaluate their proposed algorithms in a continuous time event. The evaluation scheme further limits their contributions.
Therefore, we research the intelligent virtual resource allocation, on the basis of 5G networks. We also consider the mobile virtual node so as to research the node mobility effect on the virtual resource allocation results. Instead of discussing about node mobility in an abstracted manner, we consider the concrete mobile network scenario. Unmanned aerial vehicles (UAVs) research [29], [30], [31], [32] has emerged in recent years. As UAV will play an important role in 5G era [29], we incorporate mobile UAVs into the 5G networks in order to expanding the coverage and angle of novel network services.
In this paper, we conduct a research on the virtual resource allocation in UAVs-assisted 5G networks. The formal problem model for UAVs-assisted 5G networks is involved. Another novel profit model, quantifying the effect of UAV mobility, is presented, too. With respect to the novel profit model, the penalty of virtual network service interruption is considered. Then, we propose an intelligent virtual resource allocation algorithm, labeled as Intell-UAV-5G. Our Intell-UAV-5G consists of two sub-algorithms: Intell-UAV-5G-Active and Intell-UAV-5G-Reactive. With respect to the Intell-UAV-5G-Active algorithm, it enables to predict all possible connecting access nodes and continue the implement virtual service quickly when the service interruption happens. With respect to the Intell-UAV-5G-Reactive algorithm, it can re-allocate and re-embed the virtual network immediately when the service interruption happens. In order to highlight our Intell-UAV-5G, we conduct the experiment evaluation. Experiment results demonstrate that Intell-UAV-5G outperforms the selected derived algorithms, in terms of income and virtual service acceptance. For example, the virtual service acceptance ratio advantage of Intell-UAV-5G is beyond 14% (10 000 time unit point).
(1) The new formal problem model for UAVs-assisted 5G networks is presented in this paper (Section 3). Previous problem models for solving virtual resource allocation were developed from the fixed core networks and computer networks. No researcher [7], [9], [10] had not considered integrating the UAVs into the virtual resource allocation problem model yet;
(2) Another novel profit model is proposed in Section 3. Besides of quantifying revenue and cost, the novel profit model enables to quantify the penalty. The quantified penalty results from the UAVs mobility. Previous profit models [7], [9], [10] did not consider quantifying the penalty. Not to mention quantifying the UAV mobility and service interruption;
(3) An intelligent virtual resource allocation algorithm is proposed in this paper (Section 4). The intelligent resource allocation algorithm is labeled as Intell-UAV-5G. The Intell-UAV-5G consists of two sub-algorithms. One sub-algorithm is Intell-UAV-5G-Active. The Intell-UAV-5G-Active enables to predict all possible connecting access nodes and continue the implement virtual service quickly when the service interruption happens. The other sub-algorithm is Intell-UAV-5G-Reactive, enabling to conduct the re-allocation and assignments when the service interruption happens;
(4) A comprehensive experiment is conducted in Section 5. The experiment work aims at validating Intell-UAV-5G efficiency. We select the mostly-similar resource allocation algorithms to make up the experiment simulation part. Experiments results are carefully recorded, plotted and discussed.
The rest of this paper is well organized: Related work is presented in Section 2. Problem model and profit model for UAVs-assisted 5G networks are both presented in Section 3. The intelligent algorithm Intell-UAV-5G is detailed in Section 4, including its two sub-algorithms. Experiment work and result discussion are presented in Section 5. In the end, we conclude this paper and outlook the future work.
Section snippets
Related work
Serving as the dominant issue in 5G networks, it is vital to efficiently allocate virtual resources. Efficient and optimal resource allocations per user contribute to implementing various network services and applications smoothly. In 5G networks, virtual resources required to be allocated are usually abstracted into the resource attributes of nodes and links. Multiple nodes and links form up the virtual network service. Therefore, the type of virtual resource allocation is called as VNE.
Formal problem model for UAVs-assisted 5G networks
In UAVs-assisted 5G networks, UAVs has the function of exchanging data and traffic flow with their connected data centers through a heterogeneous underlying physical network. The heterogeneous physical network usually consists of three parts: data center part, switching part, and random access part. Refer to Fig. 1. In the bottom of Fig. 1, we plot one underlying physical network. Within the data center part, it mainly consists of multiple data centers. With respect to the switching part, it
Intelligent virtual resource allocation algorithm
This section consists of four sub-sections. We firstly adopt the programming method in order to calculate the exact and optimal allocation solution per virtual network. Then, we will detail our intelligent algorithm Intell-UAV-5G. The Intell-UAV-5G algorithm consists of two sub-algorithms: Intell-UAV-5G-Active and Intell-UAV-5G-Reactive. In the next two subsections, Intell-UAV-5G-Active and Intell-UAV-5G-Reactive are presented in both Sections 4.2 Intell-UAV-5G-Active sub-algorithm, 4.3
Experiment settings
In this section, we plan to highlight our Intell-UAV-5G by conducting the experiments. With respect to the scale of underlying physical network, it has 100 physical nodes, including 5 data center nodes, 6 optical switching nodes, and 89 access nodes. All physical nodes are uniformly distributed in a two-dimensional plane of 5000*5000. Each pair of physical nodes, within its node type, has a connectivity possibility of 0.5. With respect to the storage of each data center node, it is an integer
Conclusion work
In order to expand the service coverage of 5G networks and implement the resource allocation in an intelligent manner, we research the virtual resource allocation problem in virtualized UAVs-assisted 5G networks. An intelligent algorithm Intell-UAV-5G is proposed in this paper.
The Problem model and profit model for UAVs-5G networks are firstly constructed. We also introduce the penalty function in the profit model which is the first attempt in VNE research area. Then, the Intell-UAV-5G
CRediT authorship contribution statement
Haotong Cao: Software, Validation, Writing - review & editing. Yue Hu: Investigation, Editing. Longxiang Yang: Resources, Project administration, Supervision.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
Some preliminary results [47] were partly presented in 2020 IEEE International Conference on Communications (IEEE ICC 2020), June 7–11, 2020, Dublin, Ireland. The authors would like to thank the associate editor and assigned reviewers for their excellent work and constructive comments, which improved the paper quality. The paper was partly supported by National Key Research and Development Program of China under Grant 2018YFC1314903, National Natural Science Foundation of China under Grant
Haotong Cao received the B.S. Degree in Communication Engineering from Nanjing University of Posts and Telecommunications (NJUPT) in 2015. He received the Ph.D. Degree in Communication and Information Systems from NJUPT, China, in 2020. He was a visiting scholar of Loughborough University, U.K. in 2017. He is now the Postdoc in The Hong Kong Polytechnic University, P.R. China. He has served as the TPC member of multiple IEEE conferences, such as IEEE INFOCOM, IEEE ICC, IEEE Globecom. He is also
References (47)
- et al.
Qos satisfaction aware and network reconfiguration enabled resource allocation for virtual network embedding in fiber-wireless access network
Comput. Netw.
(2018) - et al.
Providing healthcare services on-the-fly using multi-player cooperation game theory in Internet of Vehicles (IoV) environment
Digit. Commun. Netw.
(2015) - et al.
Secure clustering for efficient data dissemination in vehicular cyber–physical systems
Future Gener. Comput. Syst.
(2016) What will 5G be?
IEEE J. Sel. Areas Commun.
(2014)- et al.
UAV-empowered edge computing environment for cyber-threat detection in smart vehicles
IEEE Netw.
(2018) - et al.
The tactile internet: Vision, recent progress, and open challenges
IEEE Commun. Mag.
(2016) Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2015-2020white paper
(2016)- A. Jindal, G. Aujla, N. Kumar, SURVIVOR: A blockchain based edge-as-a-service framework for secure energy trading in...
- et al.
Data offloading in 5G-enabled software-defined vehicular networks: A stackelberg-game-based approach
IEEE Commun. Mag.
(2017) - et al.
Heuristic solution of virtual network embedding: A survey
China Commun.
(2018)
Network function virtualization: State-of-the-art and research challenges
IEEE Commun. Surv. Tutor.
A survey on algorithmic aspects of virtual optical network embedding for cloud networks
IEEE Access
A survey of embedding algorithms for virtual network embedding
China Commun.
Exact solutions of VNE: A survey
China Commun.
Multi-provider virtual network embedding with limited information disclosure
IEEE Trans. Netw. Serv. Manage.
Novel node-ranking approach and multiple topology attributes-based embedding algorithm for single-domain virtual network embedding
IEEE Internet Things J.
Novel location constrained virtual network embedding (LC-VNE) algorithms towards integrated node and link mapping
IEEE/ACM Trans. Netw.
Dynamic virtual network embedding over multilayer optical networks
J. Opt. Commun. Netw.
Dynamic embedding and quality of service-driven adjustment for cloud networks
IEEE Trans. Ind. Inform.
Secure virtual network embedding to mitigate the risk of covert channel attacks
Topology-oriented virtual network embedding approach for data centers
IEEE Access
A novel resource reservation scheme for mobile PHEVs in V2G environment using game theoretical approach
IEEE Trans. Veh. Technol.
Cited by (43)
Ubiquitous learning models for 5G communication network utility maximization through utility-based service function chain deployment
2024, Computers in Human BehaviorMADDPG-empowered slice reconfiguration approach for 5G multi-tier system
2023, Journal of Network and Computer ApplicationsResource Orchestration and Allocation of E2E Slices in Softwarized UAVs-Assisted 6G Terrestrial Networks
2024, IEEE Transactions on Network and Service ManagementBlockchain-Empowered Resource Allocation in Multi-UAV-Enabled 5G-RAN: A Multi-Agent Deep Reinforcement Learning Approach
2023, IEEE Transactions on Cognitive Communications and NetworkingMulti-Objective Optimization Approaches for Physical Layer Secure Communications Based on Collaborative Beamforming in UAV Networks
2023, IEEE/ACM Transactions on Networking
Haotong Cao received the B.S. Degree in Communication Engineering from Nanjing University of Posts and Telecommunications (NJUPT) in 2015. He received the Ph.D. Degree in Communication and Information Systems from NJUPT, China, in 2020. He was a visiting scholar of Loughborough University, U.K. in 2017. He is now the Postdoc in The Hong Kong Polytechnic University, P.R. China. He has served as the TPC member of multiple IEEE conferences, such as IEEE INFOCOM, IEEE ICC, IEEE Globecom. He is also serving as the reviewer of multiple academic journals, such as IEEE/ACM Transactions on Networking, (Elsevier) Computer Networks, IEEE Transactions on Network and Service Management, and IEEE Transactions on Network Science and Engineering. He has published multiple IEEE Trans./Journal/Magazine papers since 2016. His research interests include wireless communication theory, resource allocation in wired and wireless networks. He was awarded the Postgraduate National Scholarship of China in 2018.
Yue Hu received the B.S. Degree in Network Engineering from Nanjing University of Posts and Telecommunications in 2013, the M.S. Degree in Electronic and Communication Engineering from Nanjing University of Posts and Telecommunications, Nanjing, China, in 2016, and the M.S. Degree in Communication Engineering from Melbourne University, Melbourne, Australia, in 2015. She is currently working in China Mobile Communications Group Jiangsu Co., Ltd. Her research interests include Internet of Things technology.
Longxiang Yang is currently with the College of Telecommunications and Information Engineering, Nanjing University of Posts and Telecommunications (NJUPT), Nanjing, China. He is a Full Professor and Doctoral Supervisor of NJUPT. He is also the head of College of Telecommunications and Information Engineering, NJUPT. He has fulfilled multiple National Natural Science Foundation projects of China. He has authored and co-authored over 200 technical papers published in various journals and conferences. His research interests include cooperative communication, network coding, wireless communication theory, 5G mobile communication systems, ubiquitous networks and Internet of things.