This Small Business Innovation Research Program (SBIR) Phase I project seeks to develop a software system that transparently and dynamically offloads computations from mobile devices to the cloud for the purpose of accelerating mobile applications. The need is clear, as software requirements are already outstripping the hardware capabilities of smartphones and tablets. However, such a system has remained out of reach since current technologies cannot practically achieve the tight device-cloud coupling required for high application responsiveness, especially under the processor, battery and network constraints of modern mobile devices. Compressive replication, a novel continuous data replication method based on compressive sensing, can fill this technology gap by enabling just such a tight coupling. To elevate compressive replication to a level suitable for use in a commercial product and service, technical work will be undertaken to survey and improve its encoding/decoding mechanisms, tailor efficient network protocols to it, and evaluate its real-world performance in a prototype. The anticipated result of this feasibility study is a system that will demonstrate significant speed-up of mobile applications through transparent cloud acceleration.
The broader impact/commercial potential of this project will be a transformation of the mobile computing landscape. Application developers will be afforded a simple, principled method and a general, reusable framework for developing mobile applications that are tightly coupled to the vast resources of the cloud. In turn, this will unleash a new generation of mobile applications that rival, and even surpass, current desktop or web-based applications in sophistication, features and engagement. For end-users at large, it could mean greater productivity through more sophisticated business apps and higher quality entertainment through more immersive games. For disabled persons, cloud acceleration would make assistive user interfaces -- particularly those that combine computationally intensive natural language processing, gesture recognition, and predictive algorithms -- practical on mobile devices. The proposed system will also be a key driver of new economic activity in the mobile space, as real-time synchronization with the cloud opens up new opportunities in an extensive array of market verticals including, e.g., mobile user engagement analytics, mobile advertising, and even machine-to-machine (M2M) platform development. In short, the proposed cloud acceleration technology will have significant societal and economic impacts on many fronts.
