Research On Optimal Deployment Of Service Function Chain Based On Cloud Service Platform

From the perspective of small and medium-sized cloud service providers,this article proposes the possibility of building a distributed cloud service platform by leasing servers from large vendors such as Amazon.Unlike traditional cloud service providers building their own data centers,building data centers by leasing computing resources can greatly reduce the initial costs of cloud service providers.This platform provides users with the deployment service of the service function chain and charges fees.By arranging different types of network functions as required,a service function chain is formed.Providing a service function chain for customers in need is a market that cannot be ignored.Aiming at this new platform model,this article discusses from three aspects.First,when the cloud service provider of this platform provides users with SFC deployment services,how to achieve the maximum profit.Since the previous maximum profit algorithms have focused on self-built data centers,SFCs are often deployed on certain servers,thereby shutting down other idle servers to save electricity costs.This article discusses that in the distributed cloud scenario,computing resources and bandwidth resources are leased,and different data centers and links have different resource unit prices.This article takes into account load balancing by choosing a cheap path.And in the load balancing process,the newly arrived request is compared with the previous request of the link and the data center,thereby increasing the profit of the deployment.The simulation results show that the scheme of this article is indeed more suitable for the platform of this subject.The simulation results show that the scheme in this paper does increase the maximum profit of cloud service providers.Second,in order to build the cloud service platform of this topic,which data center and link resources should be purchased,and how much resources should be purchased,so that the hardware rental cost of the platform is the lowest.This article found that even if the platform has predicted which requests,a set of algorithms is still needed to deploy these requests at the lowest cost.Different from the general minimum cost of deploying SFC,this article attempts to centralize the deployment of requests in certain data centers and links,and obtain more discounts through large-volume purchases,thereby reducing deployment costs.This is not considered in previous SFC deployment algorithms.The simulation results prove that the algorithm in this paper does reduce the cost of leasing hardware resources.Third,this article uses engineering to simulate the cloud service platform of this subject,and realizes the process from the user initiating the deployment of SFC on the front-end page until the request is accepted and then being deployed in the target data center.Simulation process This article focuses on verifying the automated and rapid deployment of SFC,and efficient and flexible deployment is the inevitable development direction of the cloud computing era.The project implementation of this paper reduces the cost of development and testing through modularization,which also has reference significance for the construction of cloud service platforms.