Research On The Deployment Strategy Of Virtual Network Function Service Chain Based On Cloud-edge Computing In EONs

With the ongoing roll-out of the Internet-of-Things(Io Ts),the explosive growth of data enables many emerging data-driven services,such as virtual reality,autonomous driving,and e-health.By shifting processing away from remote clouds to edge locations,cloud-edge computing can enhance the quality of service(Qo S)of Io Ts and realize the expected benefits of Io Ts.However,the traditional terminal-centric static network service model has been challenging to meet the application requirements.The current rigid network architecture cannot provide flexible services for cloud-edge computing.Along with network virtualization technology development,a dynamic network service model centered on business data is emerging.Network Function Virtualization(NFV)has emerged as a promising technology to coordinate virtual network function service chains consisting of an orderly combination of virtualized network service functions.It also provides flexible services based on the bandwidth dynamic allocation technology of Elastic Optical Networks(EONs),implementing a flexible and economical alternative for network service providers to update their existing operating models.However,it is challenging for ISPs to effectively decide the deployment scheme of virtual network functional service chains and reduce network operation and maintenance costs in a cloud-edge optical network.Therefore,the paper conducts a study on optimizing the virtual network functional service chain deployment algorithm in terms of reducing network energy consumption and improving network deployment revenue,respectively.Firstly,the constraint model of deployment of the virtual network function service chain is proposed based on integer linear programming,and a series of constraints limit its deployment rules.Then,the paper establishes the energy consumption model of deployment of virtual network function service chain and the revenue model of deployment of virtual network function service chain,respectively.Secondly,an enhanced energy-aware virtual network functional service chain deployment algorithm(EEAD)based on deep reinforcement learning(DRL)algorithm is proposed to solve the deployment problem of virtual network functional service chain with low energy consumption as the optimization objective and a shortest path fitting algorithm is proposed to achieve route assignment.Finally,the paper proposed an enhanced revenue-maximizing collaborative deployment algorithm(ERMD)for virtual network function service chains to solve deploying virtual network function service chains oriented to the high revenue objective.The ERMD can achieve high deployment revenue through collaboratively deploy two phases of the virtual network functional service chain based on the interactive training of double agents.