Network Traffic Prediction in Digital Twin Network

Version	Summary	Created by	Modification	Content Size	Created at	Operation
1		Hyeju Shin	--	70	2023-10-19 10:30:37	\|
2	Fix the language issue	Hyeju Shin	+ 1532 word(s)	1602	2023-10-19 10:42:04	\| \|
3	layout & references	Sirius Huang	Meta information modification	1602	2023-10-20 03:00:15	\|

This entry is adapted from the peer-reviewed paper 10.3390/electronics12183957

Due to the spread of realistic services such as virtual reality/augmented reality, hologram content, and metaverse, communication networks are becoming more complex, and network management is becoming more complex accordingly. Digital twin network technology, which applies digital twin technology to the communications network field, is predicted to be an effective means of managing complex modern networks.

digital twin network artificial intelligence traffic prediction graph neural network data pipeline data processing

1. Introduction

With the emergence of immersive services, including virtual reality/augmented reality, hologram content, and the metaverse, communication networks are becoming increasingly complex, thereby enhancing the complexity of network management ^[1]. The digital twin network (DTN) or the network digital twin (NDT), a technology that integrates the digital twin (DT) with communication networks, has emerged as a promising new intelligent network management technology to efficiently handle these sophisticated networks ^[2]^[3]. The role of DT technology has grown in importance across a variety of fields in parallel with the advancement of simulation and computing technology ^[4]. There is active research in digital cities, including smart cities ^[5]^[6]^[7], and intelligent manufacturing technologies, including smart factories ^[8]^[9]^[10]. DT technology, which links physical space with virtual digital space, enables real-time data management based on physical models along with the full lifecycle operation and maintenance of data. This can be achieved by integrating a multi-variable, multi-scale, and multi-probability based computer simulation process to generate a virtual twin entity corresponding to the physical entity. Furthermore, it can mimic all states of the physical entity in real time, providing high fidelity and high integration. This enables the monitoring of the entity’s condition throughout the lifecycle process ^[11]. In the DTN, physical entities represent actual physical network systems that exist in the real world. These entities can be mapped to virtual twin networks in the digital domain to facilitate the comprehensive management of complex networks.

In this context, accurate modeling of the real environment in virtual reality has a significant impact on the performance of the DTN. Accordingly, various researches ^[12]^[13]^[14] on modeling traffic flow for accurate simulation, standardization, and research on DTN architecture have been conducted together ^[2]^[15]^[16]^[17]^[18]^[19]. However, the existing architectures are abstracted, making it difficult to confirm the specific data flows for the actual DTN implementation. According to ^[20], the data collection, storage, processing, and provisioning processes based on the data types should be provided for communication network management. In addition, according to ^[21], the network state data can be classified by role, and appropriate data-processing processes are required for DTN configuration.

2. Software-Defined Networks

SDN refers to the technology that separates the hardware and software functions of network equipment ^[22]. Through this separation, the hardware realm assumes the role of the data plane, transmitting the data. The software realm is divided into the control and application planes. As shown in Figure 1, the control plane can be considered to be the area where network policies are applied. It can be logically centralized and programmed via an SDN controller. This provides flexibility in network control, flow management, and congestion control. The application plane functions as the user support plane and performs policy implementation tasks.

Figure 1. SDN architecture overview.

Traditional network equipment suffers from a strong hardware dependency. This leads to difficulties in integrated management due to closed designs. An example of this issue would be the use of different custom semiconductors by different equipment vendors. SDN, on the other hand, keeps only the data transmission processing within the network infrastructure and delegates the remaining functions such as configuration, access control lists, and provisioning to a controller. This approach enables the network to be operated easily and quickly and provides flexible and efficient network management ^[23]. It combines all pieces of the network equipment into an intelligent system and solves the scalability problems intrinsic to traditional centralized structures.

SDN is an innovative paradigm in networking and offers on-demand resource allocation, easy reconfiguration, and programmable network management. The separation of the control and data planes in the network enables the flexible and consistent execution of network management and control ^[24]. As a result, the adjustment of the allocated bandwidth and paths for user services becomes dynamically manageable, facilitating traffic control and management. To this end, the SDN control plane collects real-time network status information from the data plane. It can provide a path that meets quality-of-service requirements between communicating endpoints according to user requirements.

The creation of intelligent networks has become achievable via the abstracted central controller. Network intelligence is enhanced via programmable software applications at the network control plane and the integration of AI. Recently, the emergence of the DTN and network management technology has facilitated the discussion of network automation technologies based on the SDN structure.

3. Digital Twin Network

The DTN refers to a technology that efficiently analyzes, diagnoses, simulates, and controls physical networks using the data and models within a virtual digital network that is constructed using DT technology across various network management systems and applications ^[15]^[16]. According to this definition, four key elements are required for a DTN: data, model, mapping, and interface.

The data are the foundation for building a DTN, and building an integrated data-sharing warehouse that acts as a single source of truth for the DTN facilitates the efficient storage of physical network configurations, topology, operational status data, logs, user business records, and real-time data, thereby providing data services to the network twin. The model acts as a functional source for the DTN, creating a variety of model instances via flexible combinations to provide different network applications. Mapping is required to provide a high-fidelity visualization of the physical network entities using the virtual twin network. This differentiates the DTN from a network simulation system and allows it to accurately model the state and behavior of physical network entities. The interface is a key technology for achieving physical–virtual synchronization. It connects the network service applications and physical network entities using standardized interfaces which collect and control real-time information about the physical network. This facilitates timely diagnosis and analysis. Physical–virtual synchronization is the process of updating a virtual twin entity based on the state of a physical entity. This can be achieved by collecting real-time data using an interface. Additionally, the optimized results derived from the DTN can be distributed to the physical network and controlled using the interface.

A twin network built on these four elements provides the analysis, diagnosis, simulation, and control of the physical network throughout its entire lifecycle using optimization algorithms, management methods, and expert knowledge. In the Internet Research Task Force (IRTF), the reference architecture of the DTN is presented as a structure that includes three layers and three domains, as shown in Figure 2, according to the definition of these key elements ^[15]. The three layers are the physical network layer, DTN layer, and network application layer. The three domains within the DTN layer represent the data domain, model domain, and management domain, which correspond to the data repository, service mapping model, and DTN management module subsystems, respectively. Here, the subsystem refers to one of the elements that make up the system, and it also means that it is a system in itself.

Figure 2. DTN reference architecture.

The physical network layer consists of physical network equipment that exchanges the network data and control information with the DTN layer via the southbound interface (SBI). This layer can consist of various types of physical networks, including mobile access networks, transport networks, mobile core networks, backbone networks, data centers, enterprise networks, and the industrial Internet of Things. The DTN layer consists of three subsystems: a data repository, a service mapping model, and DTN management. The data repository collects and stores a wide range of network data and provides an integrated interface to the service mapping model, facilitating the mapping of data-to-data services and models. The service mapping model completes the data-based modeling and provides data model instances for different network applications, maximizing the flexibility and programmability of network services. The network application layer communicates requirements to the twin network layer via the northbound interface (NBI) and provides services to the twin network layer using the modeled instances. After extensive validation, the twin network layer pushes control updates to the physical network equipment via the SBI.

The use of such a DTN technology enables the rapid deployment of innovative network technologies and a wide range of applications such as network operation, maintenance, optimization, network visualization, intent verification, and network self-tuning devices. This can be achieved with lower cost, higher efficiency, and greater stability compared to traditional network services. To effectively use such a DTN, the accurate modeling of the actual network as a virtual twin network must be ensured ^[25], and research is being conducted to model the future state of communication networks and data using AI prediction models ^[26]^[27]^[28]^[29].

4. Network Traffic Prediction

Network traffic prediction is becoming increasingly complex and diverse and is considered very important for network operation and management ^[30]. This is because predicting future traffic conditions (e.g., delay, traffic volume, etc.) in advance can improve the network performance. In addition, it can be used as a network analysis model to evaluate network policies before deploying network settings in the physical network within the closed loop between the DTN physical layer and the twin layer ^[12] and update the network optimization model accordingly.

Various studies based on recurrent neural network (RNN) models, which can learn temporal patterns and long-range dependencies from large-scale sequences of arbitrary length, have been conducted for traffic prediction tasks ^[31]^[32]. Recently, various studies have also been conducted using GNN models that can reflect the spatial characteristics of communication networks to improve accuracy ^[12]^[33]^[34]^[35]. However, discussions are still ongoing regarding approaches for learning large amounts of the data generated in real time in large-scale networks ^[15]^[16].

References

Nadir, Z.; Taleb, T.; Flinck, H.; Bouachir, O.; Bagaa, M. Immersive services over 5G and beyond mobile systems. IEEE Netw. 2021, 35, 299–306.
Almasan, P.; Ferriol-Galmés, M.; Paillisse, J.; Suárez-Varela, J.; Perino, D.; López, D.; Perales, A.P.P.; Harvey, P.; Ciavaglia, L.; Wong, L.; et al. Digital twin network: Opportunities and challenges. arXiv 2022, arXiv:2201.01144.
Wu, Y.; Zhang, K.; Zhang, Y. Digital twin networks: A survey. IEEE Internet Things J. 2021, 8, 13789–13804.
Chen, H.; Dang, Z.; Hei, X.; Wang, K. Design and Application of Logical Range Framework Based on Digital Twin. Appl. Sci. 2023, 13, 6589.
Raes, L.; Michiels, P.; Adolphi, T.; Tampere, C.; Dalianis, A.; McAleer, S.; Kogut, P. DUET: A Framework for Building Interoperable and Trusted Digital Twins of Smart Cities. IEEE Internet Comput. 2021, 26, 43–50.
Ghaith, M.; Yosri, A.; El-Dakhakhni, W. Synchronization-Enhanced Deep Learning Early Flood Risk Predictions: The Core of Data-Driven City Digital Twins for Climate Resilience Planning. Water 2022, 14, 3619.
Austin, M.; Delgoshaei, P.; Coelho, M.; Heidarinejad, M. Architecting smart city digital twins: Combined semantic model and machine learning approach. J. Manag. Eng. 2020, 36, 04020026.
Liu, S.; Lu, Y.; Li, J.; Shen, X.; Sun, X.; Bao, J. A blockchain-based interactive approach between digital twin-based manufacturing systems. Comput. Ind. Eng. 2023, 175, 108827.
Jwo, J.S.; Lee, C.H.; Lin, C.S. Data twin-driven cyber-physical factory for smart manufacturing. Sensors 2022, 22, 2821.
Chen, X.; Liu, F.; Ai, Y.; Xu, G.; Chen, J.; Xu, X.; Liang, W. Key characteristics analysis of industrial digital twins for smart manufacturing. Sci. Technol. Rev. 2022, 40, 45–54.
Alnowaiser, K.K.; Ahmed, M.A. Digital Twin: Current Research Trends and Future Directions. Arab. J. Sci. Eng. 2022, 48, 1075–1095.
Rusek, K.; Suárez-Varela, J.; Almasan, P.; Barlet-Ros, P.; Cabellos-Aparicio, A. Routenet: Leveraging graph neural networks for network modeling and optimization in sdn. IEEE J. Sel. Areas Commun. 2020, 38, 2260–2270.
Ferriol-Galmés, M.; Suárez-Varela, J.; Paillissé, J.; Shi, X.; Xiao, S.; Cheng, X.; Barlet-Ros, P.; Cabellos-Aparicio, A. Building a digital twin for network optimization using graph neural networks. Comput. Netw. 2022, 217, 109329.
Azzouni, A.; Boutaba, R.; Pujolle, G. NeuRoute: Predictive dynamic routing for software-defined networks. In Proceedings of the 2017 13th International Conference on Network and Service Management (CNSM), Tokyo, Japan, 26–30 November 2017; pp. 1–6.
Zhou, C.; Yang, H.; Duan, X.; Lopez, D.; Pastor, A.; Wu, Q.; Boucadair, M.; Jacquenet, C. Digital Twin Network: Concepts and Reference Architecture (draft-irtf-nmrg-network-digital-twin-arch-03). In IRTF Internet-Draft; Internet Engineering Task Force: Fremont, CA, USA, 2023.
ITU-T. Digital twin network—Requirements and architecture (Y.3090). In Telecommunication Standardization Sector of ITU; ITU-T: Geneva, Switzerland, 2022.
Hui, L.; Wang, M.; Zhang, L.; Lu, L.; Cui, Y. Digital twin for networking: A data-driven performance modeling perspective. IEEE Netw. 2022, 37, 202–209.
Wei, Z.; Wang, S.; Li, D.; Gui, F.; Hong, S. Data-driven routing: A typical application of digital twin network. In Proceedings of the 2021 IEEE 1st International Conference on Digital Twins and Parallel Intelligence (DTPI), Beijing, China, 15 July–15 August 2021; IEEE: Piscataway, NJ, USA; pp. 1–4.
Isah, A.; Shin, H.; Aliyu, I.; Oh, S.; Lee, S.; Park, J.; Hahn, M.; Kim, J. A Data-Driven Digital Twin Network Architecture in the Industrial Internet of Things (IIoT) Applications. In Proceedings of the 10th International Conference on Advanced Engineering and ICT-Convergence, AEICP, Bangkok, Thailand, 7–10 February2023.
Gartner. Rethink Network Monitoring for a Cloud Era; Gartner: Singapore, 2018.
Yang, H.; Lü, P.; Sun, T.; Lu, L.; Zhou, C. Multi-source Heterogeneous Data Processing Technology for Digital Twin Network. In Proceedings of the 2022 IEEE 22nd International Conference on Communication Technology (ICCT), Nanjing, China, 11–14 November 2022; pp. 1829–1834.
Yan, Q.; Yu, F.R.; Gong, Q.; Li, J. Software-defined networking (SDN) and distributed denial of service (DDoS) attacks in cloud computing environments: A survey, some research issues, and challenges. IEEE Commun. Surv. Tutor. 2015, 18, 602–622.
Kreutz, D.; Ramos, F.M.V.; Veríssimo, P.E.; Rothenberg, C.E.; Azodolmolky, S.; Uhlig, S. Software-Defined Networking: A Comprehensive Survey. Proc. IEEE 2015, 103, 14–76.
Kim, H.; Feamster, N. Improving network management with software defined networking. IEEE Commun. Mag. 2013, 51, 114–119.
Li, M.; Zhou, C.; Chen, D. Data Generation and Optimization for Digital Twin Network Performance Modeling (draft-li-nmrg-dtn-data-generation-optimization-00). In IRTF; Internet Engineering Task Force: Fremont, CA, USA, 2023.
Zhang, Q.; Ng, K.K.; Kazer, C.; Yan, S.; Sedoc, J.; Liu, V. MimicNet: Fast performance estimates for data center networks with machine learning. In Proceedings of the 2021 ACM SIGCOMM 2021 Conference, Online, 23–27 August 2021; pp. 287–304.
Ferriol-Galmés, M.; Paillisse, J.; Suárez-Varela, J.; Rusek, K.; Xiao, S.; Shi, X.; Cheng, X.; Barlet-Ros, P.; Cabellos-Aparicio, A. RouteNet-Fermi: Network Modeling With Graph Neural Networks. IEEE ACM Trans. Netw. 2023.
Yang, Q.; Peng, X.; Chen, L.; Liu, L.; Zhang, J.; Xu, H.; Li, B.; Zhang, G. Deepqueuenet: Towards scalable and generalized network performance estimation with packet-level visibility. In Proceedings of the ACM SIGCOMM 2022 Conference, Amsterdam, The Netherlands, 22–26 August 2022; pp. 441–457.
Shen, K.; Li, B. Learning-based Network Performance Estimators: The Next Frontier for Network Simulation. IEEE Netw. 2023, 1.
Ge, Z.; Hou, J.; Nayak, A. Gnn-based end-to-end delay prediction in software defined networking. In Proceedings of the 2022 18th International Conference on Distributed Computing in Sensor Systems (DCOSS), Los Angeles, CA, USA, 30 May–1 June 2022; IEEE: Piscataway, NJ, USA; pp. 372–378.
Vinayakumar, R.; Soman, K.P.; Poornachandran, P. Applying deep learning approaches for network traffic prediction. In Proceedings of the 2017 International Conference on Advances in Computing, Communications and Informatics (ICACCI), Udupi, India, 13–16 September 2017; IEEE: Piscataway, NJ, USA; pp. 2353–2358.
Ramakrishnan, N.; Soni, T. Network traffic prediction using recurrent neural networks. In Proceedings of the 2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA), Orlando, FL, USA, 17–20 December 2018; IEEE: Piscataway, NJ, USA; pp. 187–193.
Zebin, C.; Yichi, W.; Tang, H.; Chuanhuang, L. Research on intelligent perception model of sdn network delay. In Proceedings of the 2021 IEEE 6th International Conference on Computer and Communication Systems (ICCCS), Chengdu, China, 23–26 April 2021; pp. 452–457.
Fabien, G. Performance evaluation of network topologies using graph-based deep learning. Perform. Eval. 2019, 130, 1–16.
Yong, C.; Wei, Y.; Zhiyong, X.; Peng, L.; Zongpeng, D. Graph Neural Network Based Modeling for Digital Twin Network. IRTF, Internet-Draft. Available online: https://datatracker.ietf.org/doc/draft-wei-nmrg-gnn-based-dtn-modeling (accessed on 23 June 2023).

© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license; additional terms may apply. By using this site, you agree to the Terms and Conditions and Privacy Policy.

Upload a video for this entry

Information

Subjects: Telecommunications

Contributors MDPI registered users' name will be linked to their SciProfiles pages. To register with us, please refer to https://encyclopedia.pub/register :

Hyeju Shin

Seungmin Oh

Abubakar Isah

Ibrahim Aliyu

Jaehyung Park

Jinsul Kim

View Times: 276

Update Date: 20 Oct 2023

Table of Contents

Video Upload Options

Confirm