The Research Of Profit Optimization Model And Task Allocation Algorithm Objected To Cloud Resources

In recent years,the Internet develops rapidly,has brought earthshaking changes to the way people work and live.The development of information technology has also led to a lot of emerging industry generation,cloud computing came into being in the wave of the Internet's development.In recent years,cloud computing is becoming the mainstream of the development trend of the information industry,not only by the major Internet companies competing to use,while cloud computing resource optimization problems has become a hot academic research.The increasing scale of cloud users,more diverse,complicated and dynamic cloud task put higher requirements towards cloud resources allocation strategies.Based on the research at home and abroad,an improved cloud resource revenue optimization model,hoping to take into account the service provider's revenue,SLA agreement negotiated by the service provider and the customer and the service response time and other services quality constraints.Then a hybrid swarm intelligence optimization algorithm is used to solve the resource allocation model.It proves the validity of the model and can be solved.Finally,to analysis experiments performance of the algorithm for solving the model and task allocation strategy,the study validates high effectiveness of models and algorithms.The main contents are as follows:(1)Full account of the user on the quality of services,especially on the requirements of response time,while maximizing the benefits of cloud providers target model is established.It introduces the cloud resource management platform functional modules and the flow of information between these modules,and then estimate the response time of service requests by using the queuing theory,and set the parameters to calculate the processing time of the request.This paper considers the hierarchical user requests that a user request can be divided into multiple sub-request,assigned to a different server for processing.(2)Multiple variables constraint with each other in the resource allocation model,and the model is the nonlinear model,the solution of the model the problem belongs to NP-Hard problem.The solution of these problems may generally be Heuristic method,for example,greedy algorithm,swarm intelligence optimization algorithm,and so on.In this paper,we proposed PSO-ACO algorithm for cloud resource revenue model.PSO-ACO algorithm is a mixed group of intelligent optimization algorithm,could solve efficiently the model,less complicated than the greedy algorithm's prove process,and more faster than the converges ratio of swarm intelligence optimization algorithm,better results.In this paper,to code for the overall algorithm,and discrete variables are processed continuously.(3)Taking into account when the population size is too large or the number of requests is relatively too much,the solution of the algorithm take relatively long time in stand-alone environment,to solve the model by using MapReduce distributed parallel programming.Finally,the model is solved separately on a stand-alone environment and Hadoop-based distributed platform,the experimental results were compared and analyzed.Experiments show that,PSO-ACO algorithm can be more effective dynamic to find the optimal solution,and when the population size is relatively large,the parallel solution can save computing time and improve efficiency.