Leak detection for cloud-based IoT devices

Abstract

Web browser on a popular OS is a reasonable solution for consumer electronic applications running with cloud service on IoT platform such as interactive Kiosks or Point of Sales (POS) systems. Memory leak is more critical in those systems considering their tight memory constraints and higher standards of reliability. Since the complexity of browser architecture is beyond leak-free development cost, qualified leak detection tools are required. Still conventional tools are not capable of such heavy applications. This paper proposes an affordable leak detection mechanism to improve the performance and stability of an existing profiler. The result shows that we achieved orders of magnitude better performance than the existing UMDH.