On The Performance Optimization For The Cloud Architecture With Free Service And Registration Service

With the increasing demand for cloud computing,the types of cloud services are also increasing,and the competition among cloud providers is becoming more and more fierce.In order to appeal more users,and fostering a loyal customer group,we investigate a novel cloud architecture with free service and registration service,i.e.,free-registration cloud architecture,evaluate and optimize the system performance.Firstly,applying membership system to cloud computing system,from the perspective of forming stable customer groups,we propose a membership-based free-registration cloud architecture.Regarding the free service as the first service,the registration service as the second optional service,we can establish a multiple-server queueing model with a second optional service.Secondly,considering the problem of energy consumption caused by idle resources,a free-registration cloud architecture based on asynchronous sleep of virtual machines is proposed by introducing sleep mechanism into all virtual machines.Regarding the free service as the first service,the registration service as the second optional service,and all virtual machines sleep as all severs vacation,we can establish an asynchronous multivacation queuing model with a second optional service.Then,in order to balance the energy-saving effect of the system and the response performance of anonymous users,a free-registration cloud architecture based on asynchronous sleep of partial virtual machines is proposed by introducing a sleep mechanism into some virtual machines.Regarding the free service as the first service,the registration service as the second optional service,and part of virtual machines sleep as part of servers vacation,we can establish an asynchronous multi-vacation queuing model for part of virtual machine with a second optional service.Finally,the steady-state analysis of the queuing models are carried out by using the matrix-geometric solution method,and the system performance is evaluated from the aspects of the average response time of anonymous users and the energy-saving rate of the system.The Nash equilibrium and social optimal behavior of anonymous users are studied based on membership-based free-registration cloud architecture,free-registration cloud architecture based on asynchronous sleep of virtual machines and free-registration cloud architecture based on asynchronous sleep of partial virtual machines.We propose a pricing policy for registered VIP(Very Important Person)users to adjust the arrival behavior of anonymous users and achieve the unification of Nash equilibrium and social optimization.