A paradigm for creating cloud connected energy efficient user applications

Abstract

Applications running on smartphones are highly dependent on the user interface newline(UI) and depend on the back-end services deployed in the cloud to fetch the required newlinedata through a REST-based API. Since many applications are running on newlinesmartphones, energy consumption is significant. Saving energy on these devices newlineis beneficial not only for reducing the carbon footprint but also for end users as newlineit results in longer battery life. The paradigms, such as Model View View-Model newline(MVVM) used on the application side, require data transformations that cause newlineapplications to consume more battery. This thesis proposes a new paradigm, newlineRMVRVM (Remote-Model View Remote-View-Model), to create energy-efficient newlinecloud-connected user applications. The proposed paradigm aims to replace the newlineMVVM paradigm in developing cloud-connected UI-heavy applications that run newlineon battery-operated devices. RMVRVM paradigm moves the application logic newlineto the server-side and consequently lowers the battery consumption on devices newlinewhere the application is running. It also ensures that the data sent from the server newlineside to the device is, at most, what the application needs. Less data flows through newlinethe device s network service, resulting in better application response time and newlinelower power consumption. Moving the Models and View-Models of the application newlineto the server side makes the application code on the server side more complicated. newlineHence, it raises the question of whether a net energy saving is possible newlineusing the RMVRVM paradigm when considering the client and server-side energy newlineconsumption. We have conducted experiments and applied the RMVRVM paradigm to industrial newlinecase studies to validate its efficacy. We also selected an open-source newlineMVVM-based application and applied the migration framework to make it RMVRVM compliant. newlineWe recorded its energy consumption and response time before and newlineafter migration and compared the results.