Data as a Service (DaaS) offers an effective provisioning model able to exploit the advantages of cloud computing in terms of accessibility and scalability when data providers need to make their data available to different data consumers. Nevertheless, in settings where data are generated at the edge and they need to be propagated (e.g., Industry 4.0, Smart Cities), DaaS model suffers of some limitations: data transfer from the edge to the cloud - and viceversa - could require a significant time and privacy issues could hamper the possibility to move the data. Goal of this paper is to propose a DaaS model based on the Fog Computing paradigm, which combines the advantages of both cloud and edge computing. The proposed solution implements an adaptive multi-agent system where each agent autonomously manages the placement of data in the most convenient location considering the quality of service requirements of the user that it is serving. To guarantee the collaboration of the agents without imposing a centralized control, a reinforcement learning algorithm will be enacted to balance between the local optimum for the single data consumers and the satisfaction of the global requirements of all consumers.
Efficient Data as a Service in Fog Computing: An Adaptive Multi-Agent Based Approach
Plebani P.;Salnitri M.;Vitali M.
2022-01-01
Abstract
Data as a Service (DaaS) offers an effective provisioning model able to exploit the advantages of cloud computing in terms of accessibility and scalability when data providers need to make their data available to different data consumers. Nevertheless, in settings where data are generated at the edge and they need to be propagated (e.g., Industry 4.0, Smart Cities), DaaS model suffers of some limitations: data transfer from the edge to the cloud - and viceversa - could require a significant time and privacy issues could hamper the possibility to move the data. Goal of this paper is to propose a DaaS model based on the Fog Computing paradigm, which combines the advantages of both cloud and edge computing. The proposed solution implements an adaptive multi-agent system where each agent autonomously manages the placement of data in the most convenient location considering the quality of service requirements of the user that it is serving. To guarantee the collaboration of the agents without imposing a centralized control, a reinforcement learning algorithm will be enacted to balance between the local optimum for the single data consumers and the satisfaction of the global requirements of all consumers.File | Dimensione | Formato | |
---|---|---|---|
09944146.pdf
accesso aperto
:
Pre-Print (o Pre-Refereeing)
Dimensione
7.32 MB
Formato
Adobe PDF
|
7.32 MB | Adobe PDF | Visualizza/Apri |
11311-1226160_Plebani.pdf
accesso aperto
:
Publisher’s version
Dimensione
1.84 MB
Formato
Adobe PDF
|
1.84 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.