TY - JOUR
T1 - Optimizing Cost-Efficient SFC Routing in NTNs
T2 - A Novel Transformer-Ant Colony Optimization Framework
AU - Li, Yuanfeng
AU - Zhang, Qi
AU - Yao, Haipeng
AU - Xin, Xiangjun
AU - Zhao, Yi
AU - Gao, Ran
N1 - Publisher Copyright:
© 1967-2012 IEEE.
PY - 2025
Y1 - 2025
N2 - Non-Terrestrial Networks (NTN), including Low Earth Orbit (LEO) satellite constellations, aim to provide connectivity to remote and underserved areas. These networks now support complex network services leveraging Service Function Chaining (SFC), which sequences service functions such as firewalls and load balances to customize network behavior. In this paper, a comprehensive system model for SFC deployment in NTN is presented. The SFC deployment problem is regarded as comprising two parts: routing and VNF embedding. To improve the cost efficiency of SFC deployment in NTN, a Transformer-Ant Colony Optimization approach has been proposed. The employment of the Transformer architecture to encode SFC requests and the assistance of ACO with neural network modules in identifying optimal paths for VNF embedding are pioneered in this work. The simulation results confirm that our approach achieves superior performance in terms of cost, communication success rate, and delay, demonstrating its potential for efficient SFC deployment in NTN scenarios.
AB - Non-Terrestrial Networks (NTN), including Low Earth Orbit (LEO) satellite constellations, aim to provide connectivity to remote and underserved areas. These networks now support complex network services leveraging Service Function Chaining (SFC), which sequences service functions such as firewalls and load balances to customize network behavior. In this paper, a comprehensive system model for SFC deployment in NTN is presented. The SFC deployment problem is regarded as comprising two parts: routing and VNF embedding. To improve the cost efficiency of SFC deployment in NTN, a Transformer-Ant Colony Optimization approach has been proposed. The employment of the Transformer architecture to encode SFC requests and the assistance of ACO with neural network modules in identifying optimal paths for VNF embedding are pioneered in this work. The simulation results confirm that our approach achieves superior performance in terms of cost, communication success rate, and delay, demonstrating its potential for efficient SFC deployment in NTN scenarios.
KW - ant colony optimization
KW - low earth orbit satellite
KW - Non-terrestrial networks (NTN)
KW - service function chain
KW - transformer
UR - http://www.scopus.com/inward/record.url?scp=85214316255&partnerID=8YFLogxK
U2 - 10.1109/TVT.2024.3524583
DO - 10.1109/TVT.2024.3524583
M3 - Article
AN - SCOPUS:85214316255
SN - 0018-9545
VL - 74
SP - 8037
EP - 8051
JO - IEEE Transactions on Vehicular Technology
JF - IEEE Transactions on Vehicular Technology
IS - 5
ER -