Research And Implementation Of Distributed System For Mobile Application Detection

With the rapidly development of mobile Internet,mobile applications have gradually penetrated into all aspects of people's life.However,due to lack of effective supervision,driven by illegal profits,the security situation of mobile internet becomes more and more serious.The high-performance computing of distributed system can meet the detection needs of massive applications.Different from the traditional distributed task,mobile application detection task is difficult to split and task execution time differed greatly.There have been proposed to measure the detection task load through the complexity of the control flow,and to introduce application control flow complexity into load balancing algorithm to balance the task load between nodes in the cluster,and to improve the task execution efficiency of the distributed system.However,this strategy measures the load of the detection tasks from the code logic level,which takes a long time and the logic complexity can not accurately measure the load of the detection task.This parper starts with the characteristics of the underlying code,calculate the size of the application instructions,and measure the load of the detection task by the instruction size.The main work of this paper is as follows:1.Studied mobile application detection technology and the characteristics of mobile application detection tasks.Studied distributed task management framework and core technology points,including load balancing algorithm,task scheduling,file system,and analysis its advantages and disadvantages.2.According to the characteristics of mobile application detection tasks,starting from the more intuitive bottom-level features of mobile application,an instruction-level approach to measure the scale of mobile application is proposed.The scale of mobile application is introduced into the load balancing algorithm,and the task scheduling is guided by the load balancing algorithm.At the same time,this parper optimizes the storage of massive small files.3.Design a distributed task management system which is suitable for the characteristics of mobile application detection task,and elaborate the concrete realization of the system architecture and main function modules.4.The experiment verifies that the dynamic load balancing algorithm which introduces the scale of mobile application balances the load of detection tasks between cluster nodes more effectively and improves the execution efficiency of detection tasks.At the same time,the experiment verifies that the storage of massive small files based on TFS can effectively improve the efficiency of file reading and writing.Comprehensive experimental results,the distributed task management system designed in this paper are highly efficient and can meet the needs of mass mobile application detection.