Cloudaware: Towards Context-Adaptive Mobile Cloud Computing

The widespread use of mobile devices such as smartphones and tablets is flanked by an ever increasing supply of mobile applications. Along with this trend, expectations and requirements of users rise as well. For example, users do not want to compromise on comfortable daily routines as available on desktop computers. However, an intrinsic characteristic of mobile devices is their limited availability of resources (e.g., CPU, storage, bandwidth, energy) hindering in particular computation-intensive tasks. In this scenario, mobile cloud computing (MCC) promises to overcome these limitations by offering apparently infinite resources in the infrastructure that are transparently accessible also for mobile applications. In order to easily benefit from these offerings, dynamic code offloading has been proposed by several approaches recently. However, such solutions either do not consider the complexity arising from the dynamically changing context in mobile environments adequately or have a steep learning curve inhibiting easy adoption by developers. Therefore, this paper presents a novel approach towards context-adaptive mobile cloud computing. For that, first an extensive requirements analysis was conducted merging ISO standards with users', applications' and developers' needs. Based on this, an evaluation of related MCC-approaches allowed identifying promising concepts as well as current shortcomings. As a result, an MCC-framework, called CloudAware, is proposed that eases the development of MCC-applications by offering programming abstractions, multi-level distribution transparency, context adaptation features and is hands-free for end-users.