Edge-Computing Perspectives with Reconfigurable Hardware

Abstract

The Internet of Things is a promise of smarter technologies, with devices working together in a distributed manner, to provide more quality of service in many domains, such as industry, transports, energy, health, etc. Fog/Edge computing is probably one of the most interesting concepts as it will be a means to optimize energy and performance. However, beyond the principle, few works have really demonstrated the real potential of it, as many challenges need to be addressed at different levels: hardware design, software APIs, communication protocols, radio technologies, and so on. In this context, there is a growing interest in reconfigurable devices, as they bridge the gap between performance and energy with the advantage of the hardware flexibility. In this paper, we present a platform with sensor nodes and gateways relying on FPGAs, and showing that reconfigurable devices can be an enabling technology for edge-computing.

Energy-Efficient Non-Von Neumann Computing Architecture Supporting Multiple Computing Paradigms for Logic and Binarized Neural Networks

“…All these new computing approaches aim at performing computation directly inside the memory by exploiting novel emerging non-volatile memory (ENVM) technologies, therefore bypassing the main performance bottleneck of traditional von Neumann architectures, i.e., the communication between the memory and the processing unit over a slow bus. While operation specific hardware accelerators can achieve very high performance when executing a specific task, the possibility to reconfigure the type of operations computed in-memory may benefit resource-constrained devices for edge computing applications [11]. Among in-memory computing paradigms providing reconfigurability [4,[12][13][14][15][16], architectures based on resistive memory devices and the material implication (IMPLY) logic are a promising solution [13].…”

Eliciting Truthful Data From Crowdsourced Wireless Monitoring Modules in Cloud Managed Networks

“…For the enterprise wireless networks in unlicensed shared spectrum bands, such as the wireless networks deployed in a large university campus, an office building, and an airport, cloud managed network configuration platforms are being developed for efficient resource utilization [4]. Cloud man-aged platforms can improve their performance by utilizing real-time edge analytics of key wireless metrics, such as wireless channel utilization (CU) [5]. Recent developments in artificial neural networks (ANN) [6] have encouraged the industry and the academia to increasingly focus on utilization of not only instantaneous wireless metric values but also on the prediction of metric values.…”