Research And Implementation Of Security Resource Scheduling Based On Coral Reef Algorithm In Cloud Environment

With the rapid development of Internet technology,security technology and cloud computing are gradually intertwined to form a new security defense approach,namely,a secure shared resource pool.As a new service,"security as a service" has attracted great attention from the academic and industrial fields.Users only need to rent security resources to enjoy security services without consuming local computing resources.However,with the expansion of the secure shared resource pool,data center management costs continue to increase.Security resource pools reduce costs while ensuring security has become a concern.How to achieve the dynamic resource allocation of the secure shared resource pool is of great significance based on the changes in the resource load of the nodes over time.This thesis mainly studies the security resource scheduling strategy based on the coral reef algorithm.It focuses on the two core issues of task scheduling and resource scheduling in the secure shared resource pool.In terms of task scheduling,this thesis proposes a task scheduling method based on the coral reef algorithm.First,the application scenario of task scheduling is analyzed according to the secure cloud resource pool architecture.We propose a formal description of the task scheduling model.This model can be used to calculate load balancing rate,resource utilization and load balance stability.Then we design the coral reef representation method,coral coding scheme and optimization objective of fitness function.The matrix random mapping method is applied to improve the mutation effect of coral.Finally,we compare the completion time,convergence effect and resource load with other algorithms,including ant colony algorithm ACO,genetic algorithm GA,and polling algorithm RR.In terms of virtual machine selection,this thesis improves the algorithm based on the maximum CPU utilization selection algorithm.A virtual machine selection algorithm is proposed with two factors of CPU utilization and memory capacity.We use the number of instructions requested by the virtual machine per second and the server.The ratio of the instructions number that can be processed in one second is used as an influencing factor.The ratio of the virtual machine memory size to the server memory size is used as another influencing factor.In terms of virtual machine placement,this thesis proposes a method based on the coral reef algorithm.The coding method uses energy consumption,service level agreement SLA,and the number of migrations as evaluation indicators to construct the optimization objective of the fitness function.We compare it with DVFS and PABFD classic scheduling algorithms to analyze the effectiveness of the proposed algorithm.In terms of host state identification,this thesis proposes a method to improve the load balance of the secure resource pool.When the average host CPU utilization of the secure resource pool is higher than the upper threshold,the hardware resources are extended.When the average host CPU utilization of the secure resource pool is lower than the upper threshold,the hardware is hibernated or physical node with light load is turned off to ensure secure resources load at normal levels.Finally,a large-scale experimental simulation is performed on the algorithms proposed in this thesis with CloudSim simulation tool to verify the feasibility and superiority of the algorithm.Experimental results show that,in task scheduling,the coral reef algorithm reduces the completion time and improves resource utilization compared to the three algorithms of ACO,GA,and RR.In secure resource scheduling,the algorithm proposed in this thesis effectively reduces energy costs and solves the problem of resource scheduling in the secure shared resource pool.