Startup performance analysis and optimization of an Android banking application

Mobile banking apps are becoming more and more diffused as they allow customers to access most of the services offered by their bank using their smartphones and/or tablets. However, the need for supporting several operations and guaranteeing a high level of security to the end-users contributes to making the implementation logic of such applications more complicated. This complexity can negatively affect the application performance, thus compromising the user experience. The objective of this thesis is to identify possible bottlenecks that slow down the application startup and propose a solution to reduce the total time required to launch the app, without compromising its functionalities. The methodology followed in this work consists in employing the main tools for profiling the performance of Android applications to analyze the execution at launch time of the "Mobile Banking" app, to identify critical points. Based on the results of this analysis, the thesis proposes an alternative approach that is compliant with the Android guidelines for startup time optimization, implementing the corresponding corrective actions when deemed worth the effort (in terms of performance gain and modifications to the code base). To verify the validity of the suggested modifications, a comparison of the application launch time before and after modifying the source code has been conducted, considering multiple devices with different computational powers. The results observed showed that the proposed solutions significantly reduce the time required to launch the application both in the case of high-end, more recent devices and in the case of less-performing, older ones. Finally, the approach followed throughout this work can also be extended to other sections of the applications to conduct a deeper analysis of its performance.