Research On Profit Optimization Of Services In Cloud Computing

With the rapid development of virtualization technology,big data and artificial intelligence technology,cloud computing has been widely used in many key areas such as the Internet,government affairs,finance,transportation,medical treatment,and education.Under the cloud computing paradigm,cloud service providers(CSPs)provide users with different forms of shared services such as hardware facilities,platforms,and software through the Internet,and users can purchase and use these cloud services on demand through the network at anytime and from anywhere.As the cloud service market grows,more and more CSPs appear.Thus,in a fiercely competitive cloud service market,when facing various cloud service requests submitted by many users,how to maximize cloud service profit while improving customer experience(CX)is a key issue that CSPs need to solve.Cloud service profit optimization can be achieved by increasing cloud service revenue and reducing cloud service cost.Effective cloud service pricing strategies can increase cloud service revenue while efficient cloud platform computing resource configuration schemes and cloud service request scheduling methods can reduce cloud service cost.Existing works in cloud service profit optimization usually focus on studying CSPs but rarely explore and analyze cloud users.Although existing technologies can increase the profit of cloud services in the short term,they may cause a decline in CX due to ignorance of the true demand of customers,leading to customer churn and finally negatively affecting long-term cloud service profit growth.Regarding the above issues,this paper fully considers the characteristic and demand of cloud customers(e.g.,customer personality,customer perceived value,customer diversity,customer demand differences,customer lifetime dynamics,and customer retention rate)to design customer perception-based cloud service pricing,cloud platform computing resource configuration,and cloud service request scheduling schemes which not only maximize cloud service profit from the both aspects of increasing revenue and decreasing cost but also improve CX.Specifically,the contributions of this paper are summarized as follows.1.This paper solved the problem of cloud service pricing in the dynamic cloud service market,and proposed a user personality-guided cloud service pricing strategy.Cloud customers and the CSP are modeled respectively and on top of that,the cloud service profit optimization problem under the constraint of quality of service is formulated.To solve this problem,this paper first proposed a user personality-based customer perceived value prediction mechanism that can predict customer perceived value according to user personality,service price,and quality of service.Further,based on customer perceived value,this paper designed a dynamic cloud service pricing strategy by using reinforcement learning technique.This strategy models the cloud service pricing problem as a Markov decision process and uses the Q-learning method to solve the problem to obtain the optimal cloud service pricing decisions.2.This paper solved the problem of computing resource configuration and customer in-vestment under limited budget,and proposed a customer retention rate(CRR)-driven cloud platform computing resource configuration scheme.Cloud customers and the CSP are modeled respectively and on top of that,the cloud service profit optimization problem under the constraint of budget is formulated.To solve this problem,this paper first considered the limitation of marketing budget,and proposed a customer lifetime value-based customer investment scheme that can help the CSP invest in the valuable customers to promote its long-term profit growth.Then,this paper consid-ered the limitation of infrastructure budget,and proposed a CRR-driven computing resource investment and configuration scheme that can maximize cloud service profit while increasing CRR to promote long-term cloud service profit growth.3.This paper solved the problem of personalized cloud service request scheduling under the deadline constraint,and proposed an earliest deadline first(EDF)-based heuristic cloud service request scheduling method.Cloud customers and the CSP are modeled respectively and on top of that,the cloud service profit optimization problem under the constraint of deadline is formulated.To solve this problem,this paper proposed an EDF-based two-stage heuristic cloud service request scheduling algorithm.In the first stage,the scheduling priority of service requests is ordered according to the service request deadline while in the second stage,the algorithm intelligently assigns requests to servers for execution to maximize cloud service profit under the deadline constraint.