Service Chain Placement by Using an African Vulture Optimization Algorithm Based VNF in Cloud-Edge Computing
The use of virtual network functions (VNFs) enables the implementation of service function chains (SFCs), which is an innovative approach for delivering network services. The deployment of service chains on the actual network infrastructure and the establishment of virtual connections between VNF in...
| Autores: | , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2023 |
| País: | España |
| Institución: | Universidad de Salamanca (USAL) |
| Repositorio: | GREDOS. Repositorio Institucional de la Universidad de Salamanca |
| OAI Identifier: | oai:gredos.usal.es:10366/160209 |
| Acceso en línea: | http://hdl.handle.net/10366/160209 |
| Access Level: | acceso abierto |
| Palabra clave: | cloud-edge network VNF AVOA service function chain physical network |
| Sumario: | The use of virtual network functions (VNFs) enables the implementation of service function chains (SFCs), which is an innovative approach for delivering network services. The deployment of service chains on the actual network infrastructure and the establishment of virtual connections between VNF instances are crucial factors that significantly impact the quality of network services provided. Current research on the allocation of vital VNFs and resource constraints on the edge network has overlooked the potential benefits of employing SFCs with instance reuse. This strategy offers significant improvements in resource utilization and reduced startup time. The proposed approach demonstrates superior performance compared to existing state-of-the-art methods in maintaining inbound service chain requests, even in complex network typologies observed in real-world scenarios. We propose a novel technique called African vulture optimization algorithm for virtual network functions (AVOAVNF), which optimizes the sequential arrangement of SFCs. Extensive simulations on edge networks evaluate the AVOAVNF methodology, considering metrics such as latency, energy consumption, throughput, resource cost, and execution time. The results indicate that the proposed method outperforms BGWO, DDRL, BIP, and MILP techniques, reducing energy consumption by 8.35%, 12.23%, 29.54%, and 52.29%, respectively. |
|---|