Automated Memory Leakage Detection in Android Based Systems

Abstract

Since open platforms such as Android vary in device manufacturers and application developers, modifications in software happened in multiple layers. Therefore, every layer including OS, library, framework and application may have defects within. Especially, a memory leakage which increases memory usage and diminish overall system performance is the key issue in embedded systems with highly limited resources. In this paper, we suggest a technique that detects memory leakage by gathering memory execution information in run-time via PCB hooking and apply this technique to actual Android smartphones. The suggested technique does not require a target source code and any hardware changes for memory leakage detection, and it is characterized by maintaining the same target runtime status even in detecting while minimizing performance overhead simultaneously. We implemented an automated tool of this technique for Android Smartphone, and show that it is effective.