Cost-Oriented and Delay-Constrained Anycasting for Service Function Chain Provisioning Leveraging Cloud-Edge Collaboration in Space-Air-Ground Integrated Networks

• Published in: IEEE internet of things journal Vol. 12; no. 4; pp. 4475 - 4487
• Main Authors: Liu, Yuanhao, He, Yuxuan, Li, Yongjun, Wang, Xiang, Zhang, Kai, Ju, Min, Ma, Zhiqiang, Tian, Qin
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 15.02.2025; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Anycast; Automation; Cloud computing; cloud-edge collaboration; Collaboration; Computer architecture; Convergence; Data transmission; Edge computing; Heterogeneous networks; Heuristic algorithms; Integer programming; Internet of Things; Linear programming; Low earth orbit satellites; Mixed integer; Monitoring; Network function virtualization; network function virtualization (NFV); Nodes; Provisioning; Routing; Servers; service function chain (SFC); Space vehicles; space-air–ground integrated networks (SAGINs); Virtual networks
• ISSN: 2327-4662; 2327-4662
• DOI: 10.1109/JIOT.2024.3485640

Network function virtualization (NFV) offers a flexible and effective means to utilize heterogeneous resources for space-air-ground integrated networks (SAGINs), enabling seamless connectivity for data transmission over large spans. Converging the cloud and edge computing capabilities, SAGINs have the potential to further provision service function chains (SFCs) with various Internet applications. This is driving the need for efficient schemes of the cloud- or edge-based services in SAGINs. In this article, we propose a novel SAGIN architecture based on the cloud-serving and edge-processing collaboration. The cloud-based services are provisioned by the selected ground data centers (DCs), in which the traffic is processed by the virtual network function (VNF) hosted in the edge nodes and DCs. In such an architecture, the edge nodes enable flexible SFC provisioning solutions while DCs offer a variety of cloud-oriented network services. In addition, we apply anycast to further improve the agility of SFC provisioning. From these perspectives, we investigate the cloud-serving and edge-processing SFC provisioning problem leveraging anycast, concerning DC assignment, edge and VNF placement, SFC mapping, and delay constraints simultaneously. The joint problem is formulated by a mixed integer linear program (MILP) model to jointly minimize the communication and computation costs subject to their tradeoff. A decomposition approach is further developed for the sake of scalability. Results from numerical simulations show that the proposed approach can reduce overall costs by up to 32.99%.