Design And Implementation Of Security And Energy-saving Strategy On Android

Recent years, along with the development of network technique, communication technology and terminal technology, Mobile Internet develop rapidly. Followed by the popularization of mobile devices, a large amount of users transfer from PC to mobile terminal. As the first place mobile OS, Android continuously draws a widely attention. The open source of Android not only attracts lots of users, but also exposes more defects.Under a further research on Android native mechanism, this issue analyses and improves the defects. Also, it proposes a security enhanced framework and an energy saving optimizing system.Firstly, it analyses the application launching flow. Through the source code, the vulnerability and its destructive effect on the system is shown. Then, the acquisition and release of WackLock is discussed. A system named UnusualLock is exposed to find the abnormal ones and handle it.Secondly, the necessity of improving system permission is discussed. Based on some principles, AppDroid, a framework to enhance native security mechanism is proposed. According to its different function, AppDroid is divided into three modules:setting module, checking module and strategy saving module. Setting module is designed for interaction between users and devices. With setting module, AppDroid knows the strategy to each permission of each application. Checking module checks whether the request delivered from system server could be confirmed, ensuring each application only uses the agreed permissions. Strategy saving module saves the strategy to local disk, and recovers them to AppDroid when launching.Aiming at the deficiency of native power management, an energy saving optimizing system is achieved. Three modules is presented. Data collection module is designed to collect data using a dynamic collecting algorithm which could balances energy consumption and data accuracy. Data pre-processing module uses Decision Tree to pre-process the data and it performs better than HMM model. Decision module uses data to construct a Native Bayes Classifier and controls the energy-consuming module.In the end, the test shows that AppDroid performs well on controlling application permissions, and the energy saving optimizing system improve energy saving by 86%.