Cloud RTR: Cloud Infrastructure for Apps with Hardware

Graduate Thesis Or Dissertation

Citations

Citeable URL: https://scholar.colorado.edu/concern/graduate_thesis_or_dissertations/gf06g299p

Abstract

There has been a great deal of innovation in the software space for smart phones, however, there has been virtually no room to innovate in the architecture space. By introducing a Field Programmable Gate Array (FPGA) on mobile phone platforms, developers are able to produce hardware that their applications can use. We call re-programmable hardware on mobile phones mobware. In order to introduce mobware to mobile platforms, we leverage technology that contains a processor (ARM) coupled with FPGA so we can introduce reconfigurable logic to smart phones, expose the hardware to applications, and extend a mobile operating system to allow for software control of the current hardware configuration. There are obstacles with deploying mobware so any phone can simply download it and deploy it on their FPGA. This dynamic loading of mobware is called run-time reconfiguration (RTR). Achieving RTR of hardware requires tool support and a deployment infrastructure to efficiently implement it. We present a cloud RTR deployment model that allows for the production and distribution of mobware. The deployment models converges the phone manufacturer, the application and mobware developer, and user. With these parties converged, this allows for the sustainable production and deployment of mobware.

Creator