The Research And Optimization Of Ranking Based Cloud Service Recommendation

Today,internet technology is widely used in all walks of life,We ushered in the rapid development of cloud computing.It offers a new business computing model and service model that brings computing,storage,networking and applications together.Users pay for what they need to get the resources they need.Cloud computing has now been widely recognized by academia and major international companies and has quickly become an important research area in the field of computer science.With the popularity of cloud computing,cloud services emerge as the times require,and cloud services refer to on-demand and measurable services provided to users under the technical framework of cloud computing.However,with the rapid increase in the number of cloud services,the number of cloud services that have the same function but contain different quality of service(QoS)has also become more and more.For such a large number of cloud services,it is very difficult for users to choose a cloud service that really suits their needs.Therefore,how to recommend cloud services to users accurately and efficiently is an urgent problem to be solved.This paper studies the rankings-based cloud service recommendation,the main work is divided into the following two aspects:Firstly,by researching the existing rank-based cloud service recommendation algorithm and analyzing the deficiency,a Quantitative Service Preferences(QSPRank)algorithm based on the quantified service preference is proposed,which improves the time complexity without increasing the time complexity The Normalized Discounted Cumulative Gain(NDCG)is normalized to improve the recommended accuracy of cloud services;Secondly,the optimization of large-scale cloud service recommendation is studied.The large-scale cloud service recommendation is based on the strong computing power of big data platform because of the large amount of data,but the running time will still be relatively long.By optimizing the job scheduling strategy of the platform Shorten the running time of the task,improve the resource utilization,and then achieve the goal of optimizing cloud service recommendation.Based on the research of the existing job scheduling strategy,this paper proposes a High Response Ratio Priority Scheduler(Hrrp Scheduler),which combines the advantages of first-come-first-serve and short-job first algorithms,To a certain extent,reduced the running time and further optimized the time performance recommended by large-scale cloud services.