In the last few years we saw an increased interest in heterogeneous and reconfigurable platforms like FPGAs, thanks to their flexibility w.r.t. custom ASICs and performance w.r.t. common CPUs when taking into account specific tasks. If we focus on Internet of Things (IoT) devices and networks, FPGA boards can be leveraged as fog nodes to decentralize the workload and delegate tasks to the edges of a network, collecting in the Cloud only pre-processed data to reduce bandwidth. In this paper we propose Fog Acceleration through Reconfigurable Devices (FARD), which is a cluster of heterogeneous boards (CPU + FPGA) able to improve performance/Watt ratio, scalability and flexibility in scenarios that exploit fog computing. To this aim, in this paper we provide an accelerated fog application that leverages FARD to monitor car flows with video surveillance cameras. The proposed application leverages the PYNQ-Z1 1 platform, outperforming the software implementation running on an Intel core i7 by 3.75x in terms of execution time per frame and of 33.98x in terms of FPS/Watt.

Fog Acceleration through Reconfigurable Devices

CASASOPRA, FABIOLA;Rolando Brondolin;Marco D. Santambrogio
2019-01-01

Abstract

In the last few years we saw an increased interest in heterogeneous and reconfigurable platforms like FPGAs, thanks to their flexibility w.r.t. custom ASICs and performance w.r.t. common CPUs when taking into account specific tasks. If we focus on Internet of Things (IoT) devices and networks, FPGA boards can be leveraged as fog nodes to decentralize the workload and delegate tasks to the edges of a network, collecting in the Cloud only pre-processed data to reduce bandwidth. In this paper we propose Fog Acceleration through Reconfigurable Devices (FARD), which is a cluster of heterogeneous boards (CPU + FPGA) able to improve performance/Watt ratio, scalability and flexibility in scenarios that exploit fog computing. To this aim, in this paper we provide an accelerated fog application that leverages FARD to monitor car flows with video surveillance cameras. The proposed application leverages the PYNQ-Z1 1 platform, outperforming the software implementation running on an Intel core i7 by 3.75x in terms of execution time per frame and of 33.98x in terms of FPS/Watt.
2019
2019 IEEE 5th International forum on Research and Technology for Society and Industry (RTSI)
9781728138152
Fog computing, Reconfigurable devices, Fog acceleration
File in questo prodotto:
File Dimensione Formato  
conference_041818.pdf

Accesso riservato

Descrizione: Articolo
: Post-Print (DRAFT o Author’s Accepted Manuscript-AAM)
Dimensione 584.19 kB
Formato Adobe PDF
584.19 kB Adobe PDF   Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11311/1119785
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 3
  • ???jsp.display-item.citation.isi??? ND
social impact