Multi-access edge computing (MEC) brings data and computational resources near mobile users, with the ultimate goal of reducing latency, improving resource utilization, and leveraging context and radio awareness. Relocation policies for applications in the MEC environment are necessary to guarantee its effectiveness and performance, and can use a multitude of different data (user position and direction, availability of MEC services and computation resources, etc.). In this article, we advocate using deep reinforcement learning to relocate applications in MEC scenarios, by having MEC learn during the evolution of the system. We show the feasibility of this approach and highlight its benefits via simulation, also presenting an environment that can foster future research on this topic.

Using Deep Reinforcement Learning for Application Relocation in Multi-Access Edge Computing

De Vita F.
Primo
;
Bruneo D.
;
Puliafito A.
Penultimo
;
2019

Abstract

Multi-access edge computing (MEC) brings data and computational resources near mobile users, with the ultimate goal of reducing latency, improving resource utilization, and leveraging context and radio awareness. Relocation policies for applications in the MEC environment are necessary to guarantee its effectiveness and performance, and can use a multitude of different data (user position and direction, availability of MEC services and computation resources, etc.). In this article, we advocate using deep reinforcement learning to relocate applications in MEC scenarios, by having MEC learn during the evolution of the system. We show the feasibility of this approach and highlight its benefits via simulation, also presenting an environment that can foster future research on this topic.
File in questo prodotto:
File Dimensione Formato  
08928171.pdf

solo gestori archivio

Descrizione: Articolo principale
Tipologia: Versione Editoriale (PDF)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 615.46 kB
Formato Adobe PDF
615.46 kB Adobe PDF   Visualizza/Apri   Richiedi una copia
paper.pdf

solo utenti autorizzati

Tipologia: Documento in Post-print (versione successiva alla peer review e accettata per la pubblicazione)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 856.91 kB
Formato Adobe PDF
856.91 kB Adobe PDF   Visualizza/Apri   Richiedi una copia
Pubblicazioni consigliate

Caricamento pubblicazioni consigliate

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: http://hdl.handle.net/11570/3172890
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 6
  • ???jsp.display-item.citation.isi??? ND
social impact