Analysing the performance of mobile cross-platform development approaches using ui interaction scenarios

Abstract

For developing mobile apps, developers can choose between a native development approach, in which a unique code base needs to be maintained for each supported mobile platform, and mobile cross-platform development (MCPD) approaches. MCPD approaches allow building and deploying mobile apps for several mobile platforms from one single code base. As MCPD approaches build on top of different technologies, in this paper, we analyze the performance of MCPD approaches based on UI interactions. We developed one app natively, and two using MCPD approaches. Using automated tests, we measured CPU usage, memory consumption and rendered frames of these apps when executing UI interaction scenarios consisting of three selected UI interactions (i.e., opening/closing a navigation drawer, screen transition, and virtual scrolling). The study confirms results of previous studies showing that, compared to natively developed apps, apps developed using MCPD approaches put a higher load on mobile devices regarding CPU usage, main memory and GPU memory consumption.