Sherlockfog: Finding opportunities for MPI applications in fog and edge computing

The Fog and Edge Computing paradigms have emerged as a solution to limitations of the Cloud Computing model to serve a huge amount of connected devices efficiently. These devices have unused computing power that could be exploited to execute parallel applications. In this work we present SherlockFog...

Descripción completa

Guardado en:
Detalles Bibliográficos
Publicado: 2018
Materias:
Benchmarks
Distributed systems
Fog and edge computing
Parallel applications
Benchmarking
Fog
Computational resources
Edge computing
MPI applications
NAS parallel benchmarks
Network condition
Network topology
Parallel application
Distributed computer systems
Acceso en línea:https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_18650929_v796_n_p185_Geier
http://hdl.handle.net/20.500.12110/paper_18650929_v796_n_p185_Geier
Aporte de:
Biblioteca Digital - Facultad de Ciencias Exactas y Naturales (UBA) de Universidad de Buenos Aires
id paper:paper_18650929_v796_n_p185_Geier
record_format dspace
spelling paper:paper_18650929_v796_n_p185_Geier2023-06-08T16:29:40Z Sherlockfog: Finding opportunities for MPI applications in fog and edge computing Benchmarks Distributed systems Fog and edge computing Parallel applications Benchmarking Fog Computational resources Distributed systems Edge computing MPI applications NAS parallel benchmarks Network condition Network topology Parallel application Distributed computer systems The Fog and Edge Computing paradigms have emerged as a solution to limitations of the Cloud Computing model to serve a huge amount of connected devices efficiently. These devices have unused computing power that could be exploited to execute parallel applications. In this work we present SherlockFog, a tool to experiment with parallel applications in Fog and Edge network setups, specially focused on the MPI based applications. We propose a methodology to study feasibility of running parallel applications in Fog or Edge environments. We validate this tool contrasting experimental results with theoretical predictions reaching remarkable agreement between both. We analyze the effect of worsening network conditions for several benchmarks of the MPI version of NAS Parallel Benchmarks on fog-like network topologies. Our results show that this impact is sublinear in some cases, opening up opportunities to use distributed, increasingly ubiquitous computational resources. © Springer International Publishing AG 2018. 2018 https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_18650929_v796_n_p185_Geier http://hdl.handle.net/20.500.12110/paper_18650929_v796_n_p185_Geier
institution Universidad de Buenos Aires
institution_str I-28
repository_str R-134
collection Biblioteca Digital - Facultad de Ciencias Exactas y Naturales (UBA)
topic Benchmarks
Distributed systems
Fog and edge computing
Parallel applications
Benchmarking
Fog
Computational resources
Distributed systems
Edge computing
MPI applications
NAS parallel benchmarks
Network condition
Network topology
Parallel application
Distributed computer systems
spellingShingle Benchmarks
Distributed systems
Fog and edge computing
Parallel applications
Benchmarking
Fog
Computational resources
Distributed systems
Edge computing
MPI applications
NAS parallel benchmarks
Network condition
Network topology
Parallel application
Distributed computer systems
Sherlockfog: Finding opportunities for MPI applications in fog and edge computing
topic_facet Benchmarks
Distributed systems
Fog and edge computing
Parallel applications
Benchmarking
Fog
Computational resources
Distributed systems
Edge computing
MPI applications
NAS parallel benchmarks
Network condition
Network topology
Parallel application
Distributed computer systems
description The Fog and Edge Computing paradigms have emerged as a solution to limitations of the Cloud Computing model to serve a huge amount of connected devices efficiently. These devices have unused computing power that could be exploited to execute parallel applications. In this work we present SherlockFog, a tool to experiment with parallel applications in Fog and Edge network setups, specially focused on the MPI based applications. We propose a methodology to study feasibility of running parallel applications in Fog or Edge environments. We validate this tool contrasting experimental results with theoretical predictions reaching remarkable agreement between both. We analyze the effect of worsening network conditions for several benchmarks of the MPI version of NAS Parallel Benchmarks on fog-like network topologies. Our results show that this impact is sublinear in some cases, opening up opportunities to use distributed, increasingly ubiquitous computational resources. © Springer International Publishing AG 2018.
title Sherlockfog: Finding opportunities for MPI applications in fog and edge computing
title_short Sherlockfog: Finding opportunities for MPI applications in fog and edge computing
title_full Sherlockfog: Finding opportunities for MPI applications in fog and edge computing
title_fullStr Sherlockfog: Finding opportunities for MPI applications in fog and edge computing
title_full_unstemmed Sherlockfog: Finding opportunities for MPI applications in fog and edge computing
title_sort sherlockfog: finding opportunities for mpi applications in fog and edge computing
publishDate 2018
url https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_18650929_v796_n_p185_Geier
http://hdl.handle.net/20.500.12110/paper_18650929_v796_n_p185_Geier
_version_ 1768544294181797888

Ejemplares similares