Dissertation
| Author | Gabriel Orsini |
| Tutor | Winfried Lamersdorf |
| Title | Kontextadaptive Anwendungsarchitekturen für das mobile Cloud Computing |
| Abgabe am | 05.10.2017 |
| Abstract | The ongoing technological progress in the field of information technology has led to an ongoing miniaturization of the devices involved. This miniaturization is seen as the main enabler for the realization of Mark Weiser’s vision of Ubiquitous Computing, which describes the pervasion of everyday objects with information technology. This miniaturization has already brought about a variety of new devices, such as smartphones or tablets, which, due to their size, are characterized by enabling high mobility in conjunction with their use. These mobile devices, however, may not always meet the requirements of the users with regard to supported applications, storage capacity, and battery life, as the ongoing miniaturization is followed only partially by an opposing increase in performance. In the future, we expect increasing challenges for these mobile devices, which require the cooperation of resource-constrained mobile devices with additional resources in their environment in order to meet the users’ requirements. This cooperation is also referred to as Mobile Cloud Computing. An example is the computational offloading in natural language processing by personal assistants such as Apple Siri or Microsoft Cortana, which transfer the input of their users to a remote infrastructure for processing. In order to achieve the described cooperation, the high mobility of these devices and their frequently changing environment, i.e. their context, has to be taken into account and the interaction with the infrastructure needs to be adapted accordingly. This in turn leads to a growing need for software systems that are capable of adapting themselves independently to their respective execution contexts. A context-adaptive application is able to, for example, adapt to a changed context like a decreasing transmission bandwidth by reducing the quality, thus ensuring the usability of the application. Likewise, it can detect a low charge state of the battery and replace corresponding parts of its implementation with an energy-efficient variant in order to maximize the remaining uptime of the mobile device in battery mode. Accordingly, the main research focus of this dissertation is on the investigation and elaboration of concepts for a context-adaptive application architecture, which allows applications on mobile devices to adapt themselves by means of different adaptation strategies, for example by using the surrounding infrastructure, to enable the described scenario of distributed natural language processing by adapting the distribution strategy automatically and proactively to the current and future context of a mobile device. As a result, essential requirements for a context-adaptive application architecture have been collected, summarized in a requirements catalog, and a corresponding base architecture for mobile cloud computing has been presented. In this regard, a generic context adaptation process has been developed that allows mobile devices to predict their future context in order to adapt the behavior of a mobile application and its interaction with the infrastructure. For this purpose, different adaptation strategies have been designed which aim to extend the functionality of a mobile device, accelerate existing functionality or help to achieve a higher availability and usability of mobile applications. Based on these foundations, a concept for a context-adaptive application architecture, which allows a dynamic cooperation of mobile devices with their infrastructure, has been developed. The developed architecture enables developers to build mobile applications that are distributed and context-adaptive without having to deal with the details of the distribution and adaptation as these mobile applications independently learn and anticipate changes in their individual usage context. |
| Other formats |  |
