Fast In-Network Functionality Embedding in Software-Defined Service-Centric Networking

Joint resource allocation in integrated networking, computing, and caching frameworks has attracted plenty of attention. In software-defined service-centric networking (SDSCN), we investigate an energy cost economical functionality embedding (FE) problem. The FE problem is a non-convex quadratically...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:IEEE Global Communications Conference (Online) s. 2754 - 2759
Hlavní autoři: Li, Xiaolu, Xie, Renchao, Huang, Tao, Liu, Yunjie
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 04.12.2023
Témata:
alternating direction method of multipliers (ADMM)
Approximation algorithms
Convex functions
Costs
Functionality embedding
Global communication
Iron
non-convex quadratically constrained quadratic programming (QCQP)
Quadratic programming
Resource management
successive convex approximation (SCA)
ISSN:2576-6813
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Joint resource allocation in integrated networking, computing, and caching frameworks has attracted plenty of attention. In software-defined service-centric networking (SDSCN), we investigate an energy cost economical functionality embedding (FE) problem. The FE problem is a non-convex quadratically constrained quadratic programming (QCQP) problem which is difficult to obtain its global optimal solutions. In this paper, we propose a low-complexity high-performance algorithm for energy-economical FE design in large-scale SDSCN systems by leveraging the alternating direction method of multipliers (ADMM) together with successive convex approximation (SCA). In specific, the FE problem is first approximated as a sequence of convex subproblems via SCA. Each convex subproblem is then reformulated as a novel ADMM form to enable parallel computations and closed-form solutions. Numerical results show that our fast algorithm reduces the complexity by orders of magnitude and obtains favorable performances compared with state-of-the-art algorithms.
ISSN:2576-6813
DOI:10.1109/GLOBECOM54140.2023.10437349