Design And Implementation Of Collection And Distribution System Of Mobile APP

In recent years, with the rapid development of mobile Internet in the worldwide, the security of mobile terminals are becoming more and more tight, mobile APP malicious samples are also showed an explosive increase in the number. Mobile Internet "black chain" has been formed and developed gradually, it has caused great harm for ordinary users, operators and even the country.Therefore, National Computer Network Emergency Response Technical Team/Coordination Center of China decide to build an integrated linkage of mobile Internet malware detection and disposal "cloud platform" by integrating of existing sets of mobile Internet applications detection and disposal system. This thesis mainly implements the collection and distribution subsystems of mobile APP under the entire cloud platform based on the above background. The main work of this thesis can be summarized as the following three aspects:1) After a brief introduction to the background of thesis, we proposed system requirements and introduced the idea of load balancing for the huge number of applications. For DPri-Min-Min algorithm, each distribution needs calculating the dynamic priority of every task, so we introduce the idea of lowering new task priority instead of increasing the waiting task priority. Then we propose DPP-Min-Min algorithm. We only need calculating dynamic priority of each new task without changing the priority of the waiting task, thereby reducing the scheduling overhead in the calculation of the dynamic priority.2) By Using the method of plug-in module design system, the system are more flexible when we add new features in system and the scheduling server and download servers need only one application. By modifying the configuration file, plugins can be configured to run flexibly and servers can quickly start role-switch. System implements Min-Min algorithm, DPri-Min-Min algorithm and the DPP-Min-Min algorithm. Scheduling policy which system use can easily be set by configuration file, and new algorithms are greatly welcomed.3) At last we had experiments, the system works well and we compares DPP-Min-Min algorithm with Min-Min algorithm and DPri-Min-Min algorithm, the results show that DPP-Min-Min algorithm has a better performance respect to Min-Min algorithm in overall average waiting time of low-priority tasks, and with respect to DPri-Min-Min algorithm in the computing time overhead of scheduling. It proves that the idea is feasible to avoid repeating dynamic priority computing of waiting tasks by decreasing the priority of the new coming task.