Dynamic service function chain placement with instance reuse in Fog–Cloud Computing
The advent of Network Function Virtualization (NFV) technology has brought flexible provisioning to Fog–Cloud Computing-based Networks (FCCNs) for enterprises to outsource their network functions to data center networks. Service Function Chaining (SFC) is a networking concept in NFV by which traffic...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2023-10-01
|
| Series: | ICT Express |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2405959522001497 |
| _version_ | 1797653994911301632 |
|---|---|
| author | Xueqiang Li Cai Su Mostafa Ghobaei-Arani Mustafa Fahem Albaghdadi |
| author_facet | Xueqiang Li Cai Su Mostafa Ghobaei-Arani Mustafa Fahem Albaghdadi |
| author_sort | Xueqiang Li |
| collection | DOAJ |
| description | The advent of Network Function Virtualization (NFV) technology has brought flexible provisioning to Fog–Cloud Computing-based Networks (FCCNs) for enterprises to outsource their network functions to data center networks. Service Function Chaining (SFC) is a networking concept in NFV by which traffic is steered through an ordered set of Virtual Network Functions (VNFs) composing an end-to-end service. When hundreds of users outsource their network functions to FCCN, the optimal placement of VNFs in the network becomes important for assembling SFCs with the aim of resource utilization efficiency. Motivated by the scalability shortcomings of existing schemes, we propose Deep Reinforcement Learning (DRL)-based approaches by simultaneously considering parallelized SFC and reuse of VNFs to solve this problem, i.e., Asynchronous Advantage Actor–Critic (A3C). A parallelized SFC consists of several sub-SFCs, which can reduce delay and guarantee availability. Also, reuse of preliminary VNFs in SFC placement can improve computation acceleration. The proposed scheme pursues the maximization of the long-term cumulative reward for the trade-off between Quality of Service (QoS) and service cost. The results of the experiments show that the proposed scheme performs better than the state-of-the-art methods. |
| first_indexed | 2024-03-11T16:51:57Z |
| format | Article |
| id | doaj.art-f889039382f1473f9ab1d82ed8c7a0d9 |
| institution | Directory Open Access Journal |
| issn | 2405-9595 |
| language | English |
| last_indexed | 2024-03-11T16:51:57Z |
| publishDate | 2023-10-01 |
| publisher | Elsevier |
| record_format | Article |
| series | ICT Express |
| spelling | doaj.art-f889039382f1473f9ab1d82ed8c7a0d92023-10-21T04:22:58ZengElsevierICT Express2405-95952023-10-0195847853Dynamic service function chain placement with instance reuse in Fog–Cloud ComputingXueqiang Li0Cai Su1Mostafa Ghobaei-Arani2Mustafa Fahem Albaghdadi3International Business School, Qingdao Huanghai University, Qingdao, Shandong, 266427, China; Corresponding authors.International Business School, Qingdao Huanghai University, Qingdao, Shandong, 266427, ChinaDepartment of Computer Engineering, Qom Branch, Islamic Azad University, Qom, Iran; Production and Recycling of Materials and Energy Research Center, Qom Branch, Islamic Azad University, Qom, Iran; Corresponding authors.Information Technology Unit, Al-Mustaqbal University College, Babylon, 51001, IraqThe advent of Network Function Virtualization (NFV) technology has brought flexible provisioning to Fog–Cloud Computing-based Networks (FCCNs) for enterprises to outsource their network functions to data center networks. Service Function Chaining (SFC) is a networking concept in NFV by which traffic is steered through an ordered set of Virtual Network Functions (VNFs) composing an end-to-end service. When hundreds of users outsource their network functions to FCCN, the optimal placement of VNFs in the network becomes important for assembling SFCs with the aim of resource utilization efficiency. Motivated by the scalability shortcomings of existing schemes, we propose Deep Reinforcement Learning (DRL)-based approaches by simultaneously considering parallelized SFC and reuse of VNFs to solve this problem, i.e., Asynchronous Advantage Actor–Critic (A3C). A parallelized SFC consists of several sub-SFCs, which can reduce delay and guarantee availability. Also, reuse of preliminary VNFs in SFC placement can improve computation acceleration. The proposed scheme pursues the maximization of the long-term cumulative reward for the trade-off between Quality of Service (QoS) and service cost. The results of the experiments show that the proposed scheme performs better than the state-of-the-art methods.http://www.sciencedirect.com/science/article/pii/S2405959522001497FCCNNFVVNFSFCDRLA3C |
| spellingShingle | Xueqiang Li Cai Su Mostafa Ghobaei-Arani Mustafa Fahem Albaghdadi Dynamic service function chain placement with instance reuse in Fog–Cloud Computing ICT Express FCCN NFV VNF SFC DRL A3C |
| title | Dynamic service function chain placement with instance reuse in Fog–Cloud Computing |
| title_full | Dynamic service function chain placement with instance reuse in Fog–Cloud Computing |
| title_fullStr | Dynamic service function chain placement with instance reuse in Fog–Cloud Computing |
| title_full_unstemmed | Dynamic service function chain placement with instance reuse in Fog–Cloud Computing |
| title_short | Dynamic service function chain placement with instance reuse in Fog–Cloud Computing |
| title_sort | dynamic service function chain placement with instance reuse in fog cloud computing |
| topic | FCCN NFV VNF SFC DRL A3C |
| url | http://www.sciencedirect.com/science/article/pii/S2405959522001497 |
| work_keys_str_mv | AT xueqiangli dynamicservicefunctionchainplacementwithinstancereuseinfogcloudcomputing AT caisu dynamicservicefunctionchainplacementwithinstancereuseinfogcloudcomputing AT mostafaghobaeiarani dynamicservicefunctionchainplacementwithinstancereuseinfogcloudcomputing AT mustafafahemalbaghdadi dynamicservicefunctionchainplacementwithinstancereuseinfogcloudcomputing |