BADEP: Bandwidth and delay efficient application placement in fog-basedIoTsystems

Nowadays, fog computing is considered as a new computing method that provides the resources required to provide various services and applications to Internet of Things (IoT) devices near them. Nonetheless, in recent years, the increasing growth of IoT devices with heavily dependent on resources has...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Transactions on emerging telecommunications technologies 2021-08, Vol.32 (8), Article 4136
Hauptverfasser: Dadashi Gavaber, Morteza, Rajabzadeh, Amir
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Nowadays, fog computing is considered as a new computing method that provides the resources required to provide various services and applications to Internet of Things (IoT) devices near them. Nonetheless, in recent years, the increasing growth of IoT devices with heavily dependent on resources has made it impossible for fog computing to provide these resources efficiently in many cases, which will lead to delay in the delivery of various services. Thus, the placement of application/service in fog computing-based environments will turn into a challenging issue. This was the motivation for this paper to present a method for placement of various IoT applications in fog computing-based environments to reduce the delay in these systems. In the proposed method, it is tried to reduce the resulting delay by reducing the network usage in the fog computing layer. In doing so, the proposed method tried to place the applications so that the interdependent modules in these applications, with a high data interaction rate, are placed on a node as much as possible to reduce the need for data transfer. IFogSim simulator was used to evaluate the proposed method, where the results of the simulation show that the proposed method improved the delay by 9% and the network usage by 21% compared to the existing methods.
ISSN:2161-3915
DOI:10.1002/ett.4136