Fog Computing: Implementation of a Simple Fog Scenario Through IoT Public Services

Fog Computing:
Implementation of a Simple
Fog Scenario Through IoT
Public Services
T. Montanaro, I. Sergi, S. Limelli, L. Patrono
Department of Engineering for Innovation, University of Salento, Lecce, Italy

Bol and Split, Croatia - September 8-11, 2021 SpliTech 2021
Roadmap
❑ Introduction
❑ Goal
❑ State of the art
❑ Fog Computing requirements and scenario
❑ Architecture
❑ Implementation
❑ Discussion
❑ Conclusions and future works

Introduction
IoT: the network of devices always
connected to the Internet with the aim of
sharing information and services with other
devices and users.
Cloud Computing: acted as the de-facto
standard for tackling IoT data and services
Source: https://openclipart.org/detail/242080/internet-of-things-for-me-fossasia-2016-iot-tshirt-design-
contest

Introduction
Source: http://www.inetservicescloud.com/iot-analytics-part-1/
The more the research in
IoT domain grows, the more
the number of device
increases
And the more the
generated DATA grows

Introduction
Latency
Bandwidth
consumption
Privacy and
security
Cloud computing: main issues
Limited
customization

Introduction
To solve the problems caused by the centralized nature
of Cloud Computing the
Fog Computing has been introduced:
distributed computing, storage, control, and networking
capabilities closer to the user
Image Source:
A Mohamed, Khaled Salah , "IoT Cloud Computing, Storage, and Data Analytics", The Era of
Internet of Things: Towards a Smart World 2019 Springer International Publishing
https://doi.org/10.1007/978-3-030-18133-8_4

Goal
The present paper discusses the main characteristics of two of the most spread IoT Public
Services with the aim of demonstrating their applicability to the Fog Computing sector
Microsoft Azure
IoT Edge
AWS
IoT Greengrass

State of the art
Concentrated on the artifacts that already exploited existing platforms to develop new services to support
the development of Fog Computing solutions.
Most of the artefacts present:
• New tools for supporting the Fog Computing
• Comparison of the functionalities provided by the Cloud services (e.g., Microsoft and AWS) but without
a focus on the Fog Computing
• Comparison of the performances of Cloud services applied to Fog but without any implementation of
the analyzed use cases

Fog Computing requirements and scenario
• Fog service may be distributed anywhere along the
continuum from Cloud to Things
• The service is typically decomposed and provided by
a hierarchy of Fog Nodes (that can be any device)
• Fog Computing has to be generic enough to be
applied in several different application scenarios
Image Source:
M. De Donno, K. Tange and N. Dragoni, "Foundations and Evolution of Modern
Computing Paradigms: Cloud, IoT, Edge, and Fog," in IEEE Access, vol. 7, pp. 150936-
150948, 2019, doi: 10.1109/ACCESS.2019.2947652

The scenario is set in a hospital:
• Ambient temperature monitoring of a room that contains bags of blood ->
never exceeded 6° C
• If temperature > 6° C : visual local alarm + alert message
• If temperature < 0° C : data are not sent
• If Internet connection is absent: locally buffered data to be, then, sent when
the connection is restored
Fog Computing requirements and scenario

Architecture
A simple architecture was
designed to support the running
"hello world" Fog scenario

Implementation
Raspberry Fog Node
Raspberry Sensor Cloud platform
Microsoft Azure
IoT Edge
AWS
IoT Greengrass

Implementation
Microsoft Azure
IoT Edge
AWS
IoT Greengrass
• "IoT Hub" -> Cloud service
• "IoT Edge" -> Fog Node
• "Downstream Device" -> Sensor
• "AWS IoT Core" -> Cloud service
• "Greengrass Core Device" -> Fog Node
• "AWS IoT Core" SDK -> Sensor

Discussion
• Docker Containers to isolate one program from
each other
• Code development and deployment simplified by
dedicated GUIs
• Security routes have to be defined in setup
phases, otherwise no message or data can be
exchanged
• The code runs directly on the device
• Easy interaction with device hardware resources
• Security routes has to be defined in setup
phases, otherwise no message or data can be
exchanged
Microsoft Azure
IoT Edge
AWS
IoT Greengrass

Conclusions and Future works
Conclusions
• An analysis of the main functionalities and tools exposed by the two providers was presented to highlight
which of them can be actually exploited in the Fog domain
• Demonstrated that 2 of the most spread Cloud service providers expose services that can be exploited to
implement Fog Scenarios, although none of the providers presents such possibility in its documentation
Future works
• Analysis of other important Cloud platforms
• Use the simple architecture presented as a basis to realize more complex prototypes
• Exploit the running "hello world" scenario as a basis to demonstrate the applicability of self-made software
in the Fog domain

Thanks for your attention!
teodoro.montanaro@unisalento.it
ilaria.sergi@unisalento.it
stefano.limelli@studenti.unisalento.it
luigi.patrono@unisalento.it
T. Montanaro, I. Sergi, S. Limelli, L. Patrono
Department of Engineering for Innovation, University of Salento, Lecce, Italy

Fog Computing: Implementation of a Simple Fog Scenario Through IoT Public Services

More Related Content

What's hot (20)

Similar to Fog Computing: Implementation of a Simple Fog Scenario Through IoT Public Services (20)

Recently uploaded (20)

Fog Computing: Implementation of a Simple Fog Scenario Through IoT Public Services