Research On Resource Allocation And Optimization Mechanisms For Cloud Computing Systems

Cloud computing greatly changes the affording mode of the service,and the computing services can be paid on demand like flow.In order to improve the utilization efficiency of limited computing resource,and reduce the service providers' s cost of operation and promote the satisfaction of the customers,how to allocate resources rationally and the efficiently use the computational resources,storage resources and network resources has become an urgent problem.Facing a large number of cloud task requests,drawing up a reasonable and effective plan of resource allocation is the basis for optimal utilization of resources.The traditional cloud computing resource allocation scheme only considers the completion time of the cloud tasks.But the real demand is multiple,like load balancing and so on,those elements also have important position.Therefore,this paper proposes a cloud resource allocation algorithm which based on genetic-bacterial foraging algorithm,and consider the load of different resource nodes(virtual machine),take the completion time and load balancing of the tasks as a constraint to the allocation of cloud resources.At present,the commonly used algorithms for solving cloud resource allocation are some heuristic algorithms such as simulated annealing algorithm,particle swarm optimization algorithm,colony algorithm and genetic algorithm.However,traditional single algorithm often has various shortcomings.Such as initial pheromone accumulation of pheromone algorithm which has longer relative time.It doesn't has any feedback information about System in the search of the genetic algorithm,it may cause the bind search and the low efficiency of the last part of the algorithm.Therefore this paper introduces a bacterial foraging algorithm to solve the blind search of genetic algorithm,which may lead to partial optimal problem.A hybrid optimization algorithm based on genetic-bacterial foraging is proposed to realize the allocation and scheduling of the cloud resource.Firstly,in this paper,I completed the improvement of the initial population of the traditional genetic algorithm,the optimization of the individual similarity judgment,and enriched the diversity of the population.Then two constraint conditions,the task completion time and load balancing,are added to the adaptation function in the process of cloud resource allocation.At last,for improving the adaptive probability,I set up to the crossover and mutation operators.Thus the convergencerate increased,and it avoided some of the situation into the partial optimal.Secondly,the genetic algorithm is introduced into the genetic algorithm,and the optimal solution of the genetic algorithm is used as the initial bacterial position distribution of the bacterial foraging algorithm.Then,the optimal solution is obtained by the information from the bacteria feeding algorithm.Finally,I completed the algorithm and cloud resource allocation simulation with The CloudSim.The experimental results show that the cloud task completion time is lower and it has the ability of the load balancing,and the hybrid algorithm is superior to the traditional genetic algorithm in the convergence rate.