Heterogeneous Secure Multi-level Remote Acceleration Service for Low-power Integrated Systems and Devices

• DOI: 10.1016/j.procs.2016.08.287
• Publication Date: 2016-10-17T05:18:10Z
• Authors: Lara López; Francisco Javier ...; Terpsichori-Helen...; Sokol Kosta; Cheol-Ho Hong; Raffaele Montella; Iakovos Mavroidis; Carles Fernández
• Keywords: accelerators; Cloud computing; devices; HPC; offloading; QoS; Computer Science (all); 000; QoS; 02 engineering and technology; 004; accelerators; cloud computing; devices; HPC; offloading; QoS; Computer Science (all); accelerators; HPC; offloading; 0202 electrical engineering, electronic engineering, information engineering; Cloud computing; devices

This position paper presents a novel heterogeneous CPU-GPU multi-level cloud acceleration focusing on applications running embedded systems found low-power devices. A runtime system performs energy and performance estimations in order to automatically select local CPU-based GPU-based tasks that should be seamlessly executed more powerful remote devices or infrastructures. Moreover, it proposes, for the first time, secure unified model where almost any device infrastructure can operate as an accelerated entity and/or accelerator serving other less way.