| The multi-mode electric Internet of Things(IoT)adopts advanced,reliable,stable and efficient multi-mode communication technology and system to achieve full coverage of power communication supporting secure,reliable and flexible access for services in electric IoT.With the continuous growth of electric IoT service traffic,it is necessary to continuously deploy and update network device to satisfy the requirements of diversified electric IoT service.However,the realization of traditional network functions requires dedicated network elements,and the high coupling of hardware and software cannot meet the communication service quality requirements of delay-sensitive services in electric IoT.Softwaredefined networking(SDN)and network function virtualization(NFV)provide a way to realize service orchestration in electric IoT.In this paper,we firstly consider the heterogeneity of communication media and electromagnetic interference in multi-mode electric Internet of Things,and construct an SDN/NFV-based service orchestration architecture in multi-mode electric Internet of Things to meet the customization requirements.Secondly,the virtual network function(VNF)embedding problem in service function chain(SFC)is modeled as minimizing the total service delay by two binary variables of optimizing the VNF embedding and virtual link mapping under certain constraints.Thirdly,the optimization problem is transformed into a multi-stage many-to-one matching problem among VNFs and physical nodes,and is solved by a pricing matching-based multi-phase many to one VNF embedding(PMVE)algorithm to minimize the total service delay,where at each phase the matching between the VNF set with the same sequence on different SFCs and the physical node set is achieved.Finally,the feasibility and effectiveness of the proposed PMVE algorithm is validated through case studies.The simulation results show that compared with a matching-based VNF scheduling algorithm and a global resource capacity-based heuristic algorithm,PMVE can reduce the total delay of services by 22.09%and 39.53%,and ensure the delay-sensitive service requirements in electric IoT. |
