海量移动应用检测任务的分布式调度技术研究与实现

发布时间：2018-06-04 16:35

本文选题：应用检测 + 分布式调度　；参考：《北京邮电大学》2016年硕士论文

【摘要】：随着移动互联网的快速发展,全网移动应用数量逐年攀升,人们在生活中越来越多地依赖移动智能设备,不法商家找到可乘之机,恶意移动应用层出不穷,移动互联网的安全形势也愈发严峻。现有的移动应用检测方法已经无法满足大批量应用快速检测的需求,因此,本文针对海量移动应用快速检测需求,对海量移动应用检测任务的分布式调度技术进行了深入研究,提出了基于应用控制流复杂度的分布式调度解决方案,设计并实现了一个海量移动应用检测任务分布式调度系统。本文首先提出了海量移动应用检测任务的应用场景和具体业务需求。随后分析了移动应用检测的现状并提出分布式系统在海量移动应用检测问题的应用可行性,同时通过对移动应用及其检测任务特点的研究,提出了应用控制流复杂度与检测效率之间存在线性关系。随后,本文对移动应用检测任务分布式调度技术中的关键问题如动态负载均衡算法、任务切分及基于关键路径的表调度技术等分别展开了讨论和研究,并引入移动应用控制流复杂度作为优化算法的一个重要衡量参数。最后,本文设计了一个海量移动应用检测任务的分布式调度系统,该系统分为客户端、中心管理节点、移动应用检测任务处理节点、监控容错模块、数据持久层、通信模块,本文分别阐述了对上述六个模块的设计与实现。本文通过实验论证了对上述海量移动应用检测任务的分布式调度系统的可行性,通过实验中的功能评估和性能评估,验证了海量移动应用检测任务的分布式调度技术可以有效提高移动应用检测效率,满足大规模批量快速检测的业务需求。
[Abstract]:With the rapid development of the mobile Internet, the number of mobile applications in the whole network is increasing year by year. People rely more and more on mobile intelligent devices in their lives. Illegal merchants find the opportunity to take advantage of it, and malicious mobile applications emerge in endlessly. The security situation of mobile Internet is becoming more and more serious. The existing mobile application detection methods can not meet the needs of mass application fast detection. Therefore, the distributed scheduling technology of mass mobile application detection task is deeply studied in this paper, aiming at the demand of mass mobile application rapid detection. A distributed scheduling solution based on the complexity of application control flow is proposed, and a distributed scheduling system for mass mobile application detection tasks is designed and implemented. In this paper, the application scenario and specific business requirements of mass mobile application detection task are proposed. Then it analyzes the current situation of mobile application detection and puts forward the feasibility of distributed system in mass mobile application detection. At the same time, the characteristics of mobile application and its detection task are studied. A linear relationship between the complexity of application control flow and detection efficiency is proposed. Then, the key issues in distributed scheduling of mobile application detection tasks, such as dynamic load balancing algorithm, task segmentation and critical path-based table scheduling technology, are discussed and studied in this paper. The complexity of mobile application control flow is introduced as an important parameter of optimization algorithm. Finally, this paper designs a distributed scheduling system for mass mobile application detection tasks. The system is divided into client, central management node, mobile application detection task processing node, monitoring fault tolerance module, data persistence layer, communication module. This paper describes the design and implementation of the above six modules. In this paper, the feasibility of the distributed scheduling system for the massive mobile application detection tasks mentioned above is demonstrated through experiments, and the function evaluation and performance evaluation in the experiment are carried out. It is verified that the distributed scheduling technology of mass mobile application detection task can effectively improve the efficiency of mobile application detection and meet the needs of large-scale batch rapid detection.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP311.52

【参考文献】