6G Network AI Architecture for Everyone-Centric Customized Services

Yang, Yang; Ma, Mulei; Wu, Hequan; Yu, Qin; Zhang, Ping; You, Xiaohu; Wu, Jianjun; Peng, Chenghui; Yum, T.-S.P.; Aghvami, Hamid; Li, Geoffrey Y; Wang, Jiangzhou; Gao, Peng; Tang, Xizi; Cao, Changyong; Thompson, John F.; Wong, Kai‐Kit; Chen, Shanzhi; Debbah, Mérouane; Dustdar, Schahram; Eliassen, Frank; Chen, Tao; Duan, Xiangyang; Sun, Shi‐Gang; Tao, Xiaofeng; Zhang, Qinyu; Huang, Jianwei; Cui, Shuguang; Zhang, Wenjun; Li, Jie; Gao, Yue; Zhang, Honggang; Chen, Xu; Ge, Xiaohu; Xiao, Yong; Wang, Cheng-Xiang; Zhang, Zaichen; Ci, Song; Mao, Guoqiang; Li, Changle; Shao, Ziyu; Zhou, Yong; Liang, Junrui; Li, Kai; Wu, Liantao; Sun, Fanglei; Wang, Kunlun; Liu, Zening; Yang, Kun; Wang, Jun; Gao, Teng; Shu, Hongfeng

doi:10.48550/arxiv.2205.09944

Cited by 6 publications

(5 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We target a DGPE service upon a multi-tier architecture as in Fig. 1, where edge servers are geographically distributed in different areas and clients have their data individually collected from certain access points (e.g., 5G base station, IoT gateway) [71], [72]. For ease of formulation, we focus the GNN processing workload on inference.…”

Section: A System Overviewmentioning

confidence: 99%

GNN at the Edge: Cost-Efficient Graph Neural Network Processing over Distributed Edge Servers

Zeng¹,

Chongyu²,

Huang³

et al. 2022

Preprint

View full text Add to dashboard Cite

Edge intelligence has arisen as a promising computing paradigm for supporting miscellaneous smart applications that rely on machine learning techniques. While the community has extensively investigated multi-tier edge deployment for traditional deep learning models (e.g. CNNs, RNNs), the emerging Graph Neural Networks (GNNs) are still under exploration, presenting a stark disparity to its broad edge adoptions such as traffic flow forecasting and location-based social recommendation. To bridge this gap, this paper formally studies the cost optimization for distributed GNN processing over a multitier heterogeneous edge network. We build a comprehensive modeling framework that can capture a variety of different cost factors, based on which we formulate a cost-efficient graph layout optimization problem that is proved to be NP-hard. Instead of trivially applying traditional data placement wisdom, we theoretically reveal the structural property of quadratic submodularity implicated in GNN's unique computing pattern, which motivates our design of an efficient iterative solution exploiting graph cuts. Rigorous analysis shows that it provides parameterized constant approximation ratio, guaranteed convergence, and exact feasibility. To tackle potential graph topological evolution in GNN processing, we further devise an incremental update strategy and an adaptive scheduling algorithm for lightweight dynamic layout optimization. Evaluations with real-world datasets and various GNN benchmarks demonstrate that our approach achieves superior performance over de facto baselines with more than 95.8% cost reduction in a fast convergence speed.

show abstract

Section: A System Overviewmentioning

confidence: 99%

GNN at the Edge: Cost-Efficient Graph Neural Network Processing over Distributed Edge Servers

Zeng¹,

Chongyu²,

Huang³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…To achieve adaptability across various scenarios, Service-Based Architecture (SBA) [2] leveraging software and virtualization technology is crucial. In the forthcoming 6G era, the integration of Artificial Intelligence (AI) technology will enable an autonomous and intelligent SBA, facilitating ubiquitous AI as a Service by autonomously managing and allocating resources and functions in cloud-edge-end networks [3][4][5][6].…”

Section: Introductionmentioning

confidence: 99%

An Autonomous Deployment Mechanism for AI Security Services

Wang,

Zhou,

et al. 2024

IEEE Access

View full text Add to dashboard Cite

Future network architectures are expected to be autonomous, intelligent, and service-based, posing new security challenges. To address these challenges, the Artificial Intelligence (AI) security service emerges as a promising solution. However, the complex service configurations and performance guarantees hinder the autonomous deployment of the AI security service. This paper proposes an autonomous deployment mechanism in Software-Defined Networking/Network Function Virtualization (SDN/NFV) enabled networks. First, our mechanism introduces user and decision planes on top of the control plane, enabling hierarchical intent expression and translation from user security intent to security policies. Then, we analyze the embedding problem of the AI-based Security Function Chain (AISFC) during security policy generation. We formulate the AISFC embedding problem as an Integer Linear Programming (ILP) task to minimize the total response delay. By decomposing it into AISF placement and routing, we design a heuristic algorithm with polynomial time complexity. Finally, we validate the proposed mechanism through a prototype system and numerical simulations, demonstrating its ability to autonomously translate, implement, and guarantee the user security intent. Comparative analysis shows that our approach considering the relationship between available computing resources and delay achieves smaller response delays than the baseline. Furthermore, our algorithm achieves a gap from optimality approximately 28.57% smaller than the greedy algorithm and supports networks that are 4.34 times larger in scale than the exact solution within a 2-second execution time.

show abstract

“…This is because unlike 5G networks that aim to improve the network performance (e.g., peak data rate and service coverage), 6G networks will be able to tailor customized services to guarantee everyone's quality of experience (QoE). To achieve this goal, the 6G betwork AI architecture needs to utilize data from every user's device, and integrate heterogeneous network resources and ubiquitous intelligence from the cloud to the edge [5]. Although such a cloud-edge 6G Network AI Architecture can natively incorporate FL to support user-centric AI, the individual customized and multi-dimensional service requirements will bring critical heterogeneity challenges.…”

Section: Introductionmentioning

confidence: 99%

Optimization Design for Federated Learning in Heterogeneous 6G Networks

Luo¹,

Ouyang²,

Sun³

et al. 2023

Preprint

View full text Add to dashboard Cite

With the rapid advancement of 5G networks, billions of smart Internet of Things (IoT) devices along with an enormous amount of data are generated at the network edge. While still at an early age, it is expected that the evolving 6G network will adopt advanced artificial intelligence (AI) technologies to collect, transmit, and learn this valuable data for innovative applications and intelligent services. However, traditional machine learning (ML) approaches require centralizing the training data in the data center or cloud, raising serious user-privacy concerns. Federated learning, as an emerging distributed AI paradigm with privacy-preserving nature, is anticipated to be a key enabler for achieving ubiquitous AI in 6G networks. However, there are several system and statistical heterogeneity challenges for effective and efficient FL implementation in 6G networks. In this article, we investigate the optimization approaches that can effectively address the challenging heterogeneity issues from three aspects: incentive mechanism design, network resource management, and personalized model optimization. We also present some open problems and promising directions for future research.

show abstract

6G Network AI Architecture for Everyone-Centric Customized Services

Cited by 6 publications

References 0 publications

GNN at the Edge: Cost-Efficient Graph Neural Network Processing over Distributed Edge Servers

GNN at the Edge: Cost-Efficient Graph Neural Network Processing over Distributed Edge Servers

An Autonomous Deployment Mechanism for AI Security Services

Optimization Design for Federated Learning in Heterogeneous 6G Networks

Contact Info

Product

Resources

About