Collection And Application Of Information Of Retailers In Nation Under High Concurrency

Up to now,the numbers of user that the enterprise face to has outstripped nine million,with the deepening of informatization and datamation that collecting user data and information is imperative.In the face of a large number of users,the number of visits to the server had risen sharply and the amount of data was explosive increased.At current status,the servers and other hardware devices have been unable to be satisfied with the high-concurrency state and demand of big data.If the enterprise just increase the server cluster composed of several servers to improve performance of hardware and software,to solve problems of the large number of network requests and data,which is obviously not a good solution.By analyzing and comparing the traditional data collecting mode and business type,this paper designs a new data collecting mode based on mobile devices.In order to meet the demands of enterprise and users,this paper analyzes the problems of the high concurrency and large data processing in the new collecting mode,and make the further studies on the load balancing scheduling strategy.Based on the features of high concurrency and scheduling algorithms,this paper raise a dynamic load balancing algorithm which is based on task prediction,contrasting of the multi-time slice scheduling mechanism.The algorithm is further adjusted and improved by analyzing the multi-dimensional Markov chain and queuing theory.The experimental results show that the algorithm can improve the performance and load of the whole system.The specific research contents are as follows:1)In this paper,a new data collection model based on mobile terminal is designed according to the business type and data characteristics of the current enterprise by comparing the traditional data collection model.This model turns APP and WeChat into the entrance of data collection and builds the server cluster to perform high-level concurrent service scheduling.The data center was created using the strategy based on the Hadoop platform,the data was analyzed and processed on it.The business model and the nature of the data submitted by the national users who are optimized.In addition to this the model is optimized preliminarily.By analyzing the hierarchical structure and technical points in the model system,the problem of how to efficiently schedule the task requests and adjust the load of the server node in the high concurrency environment is put forward and solved in this paper.2)In-depth study of CPU and MEM based scheduling algorithm,based on the characteristics of the actual task request,improved and proposed based on the forecast mechanism load balancing scheduling algorithm.Each server node collects the load of other nodes and forecasts the service type and arrival rate of the network request,dynamically adjusts the request distribution,reduces the waiting time of the request and shortens the idle time of the server to achieve the effective utilization of resources.The overall load of the system reaches the equilibrium state.Experimental results show that the algorithm has better performance in shortening the response time and better performance than CPU and MEM based scheduling algorithm.3)According to the deficiency of load balancing algorithm based on prediction mechanism,using the knowledge of queuing theory,the load of network server is optimized,and the waiting,processing and suspending of task request are arranged reasonably.By analyzing the multi-dimensional Markov mechanism,the features and links of the subsequent network request are predicted.Based on the multi-time slice polling strategy,the scheduling model based on Markov prediction and queuing model is proposed.Through the verification and analysis of the experiment,the model can coordinate the load status of each server node,distribute the subsequent requests and reduce the response time.4)According to the data collection model and the load balancing technology,a set of data collection system of mobile terminal data acquisition platform,business processing platform and data processing platform is set up.And load balancing technology is applied to the data processing,speed up the server processing time,which demonstrated the data acquisition system design and implementation of the results.