Internet of Things Architectures: A Comparative Study

Marcela G. dos Santos1a , Darine Ameyed2b , Fabio Petrillo1c , Fehmi Jaafar3d , Mohamed Cheriet2e
1 Dpartement
de Informatique et mathmatique, Universit du Qubec Chicoutimi, Chicoutimi, Qubec, Canada
2 cole
de Technologie Suprieure, University of Quebec, Montreal, Qubec, Canada
3 Computer Research Institute of Montral, Montreal, Qubec, Canada
a [email protected], b [email protected], c,e darine.ameyed.1, [email protected], d [email protected]

Keywords: Internet of Things, IoT, Architectures, Layers-Model, Providers

arXiv:2004.12936v1 [cs.SE] 27 Apr 2020

Abstract: Over the past two decades, the Internet of Things (IoT) has become an underlying concept to a variety of
solutions and technologies that it is now hardly possible to enumerate and describe all of them. The concept
behind the Internet of Things is as powerful as it is complex, and for the components in the IoT solution to
mesh together perfectly, they all have to be part of a well-thought-out structure. That is where understanding
the IoT architecture becomes paramount. Because of the vast domain of IoT, there is no single consensus
on IoT architecture. Different researchers and organizations proposed different architectures under a variety
of classifications, mainly: conceptual, standard and, industrial or commercial adoption. It is indispensable
to make a systematic analysis of IoT architecture to be able to compare the industrial proposals and identify
their similarities and their differences. In this work, we summarize information about seven IoT industrial
architectures in order to propose an approach that make possible a comparative analysis between different IoT
architectures. This work presents two main contributions: (i) an approach for analyzing and comparing IoT
architectures using Layer-Model; (ii) a comparative study of seven industrial IoT architectures.

1 INTRODUCTION ferent IoT architectures. Moreover, Pratap Singh et

al. (Pratap Singh et al., 2020) affirmed that there is in
The Internet of Things (IoT) marketplace is growing general three different ways to classify the IoT archi-
spectacularly in the last few years. Indeed, Huawei tectures: domain-specific architectures, layer-specific
expectation for the number of devices connected is architectures, and industrial or commercial defined ar-
100 billion by 2025. Besides that, the impact of In- chitectures.
ternet of Things Market on the global economy is In the industrial context, we noticed that there is
enormous. McKinsey Global Institute expected this a variety of IoT architecture used and presented (Fig-
impact to be around 10 trillion US dollars by 2025 ure 1). Thus, it is indispensable to make a system-
(Ahmed et al., 2019). atic analysis of IoT architecture based on a reference
Nevertheless, there are many challenges in devel- model to be able to compare the industrial proposals
oping IoT applications: the lack of general guidelines and identify their similarities and their differences.
or frameworks that handle low level communication Thus, in this paper, we selected and analyzed
and simplify high level implementation, the using of seven industrial architectures in accordance of one
multiple programming languages to implement IoT of the reference model of IoT architectures, i.e. the
applications, the diversity of the communication pro- Layer-based architecture model.This study presents
tocols, and the high complexity of distributed com- two main contributions: (i) an approach for com-
puting (Ammar et al., 2018). paring IoT architectures in according with a reference
Several researchers pointed out the using of IoT IoT architecture and (ii) a comparative study of seven
Architecture as a tool to clarify the complexity of the of the main industrial IoT architectures (Intel, Mi-
IoT solutions and to provide a better comprehension crosoft, Cisco, Google,and IBM, Ericsson, Amazon).
of the issues that may threaten them (Alshohoumi The audience of this study are (i) researchers in-
et al., 2019). terested on modeling IoT architectures, and (ii) prac-
Concretely, Alshohoumi et al. (Alshohoumi et al., titioners interested on comparing industrial IoT archi-
2019) analyzed 148 studies and identified sixteen dif- tectures.
 varying from the basic to the end-to-end solutions.
The early IoT model was a three-layer architec-
ture presented by Gubbi et al. (Gubbi et al., 2013).
Basically, it consists of perception as a ground layer
including sensors and actuators as things, cloud as an
information processing layer, and application layer
that allows users’ interaction as the third layer. Ex-
tended by adding a business layer to provide the four-
layer model (Muccini and T. Moghaddam, 2018).
Furthermore, A. Al-Fuqaha et al. (Al-Fuqaha
(a) Cisco Architecture (Cisco, 2014) et al., 2015) gives the definition of IoT architecture
as middleware layer-based and five-layer model. The
middleware layer-based, include service composition,
service management and object abstraction. Besides,
there is the six-layer model adding fog layer or a gate-
way layer to the five-based model including also edge
and hybrid edge-cloud (Pan and McElhannon, 2018).
Finally, the recent proposal for the IoT layered
architecture is delivered by Cisco as a seven-layer
model. The previous architecture was changed by
adding a user and process layer and edge computing
layer (Cisco, 2014).
(b) Microsoft Architecture (Microsoft, 2018)

Figure 1: Examples of IoT Architectures

2.1 IoT Layers-Model: 3,5 and 7-Layers
The remaining paper is organized as follows: Sec- Model
tion 2 provides an overview about the IoT Architec-
tures. Section 3 presents the related work. Section 4
is the kernel of the work, which provides a approach There are many model Layers Model Architectures
to analyze and model IoT architectures using Layers- described by the literature. For example, in the sys-
Model. Section 5 presents the modeling of seven in- tematic review about the Internet of Things architec-
dustrial architectures using our approach. Section 6 ture, Alshohoumi et al. (Alshohoumi et al., 2019) no-
performing the analyze of seven industrial architec- ticed, after analyzing 148 studies with 16 mapping
tures and the lessons learned in our study. Section 7 architectures, that the most of architectures can be
explains the threats to validity. Finally, Section 8 syn- classified as a three, four or five layers. Another sur-
thesizes the final remarks and future work. vey about IoT architecture is the study conducted by
Pratap Singh (Pratap Singh et al., 2020) noticed that
the model layer classification applied is three, four,
2 IoT LAYER MODEL five, six, seven layers.
In our current study, we have focus on the 7-
While acknowledging the existence of other IoT Layers architecture model because it is considered the
architectures classification as conceptual, domain- complete model, regarding the complexity of the IoT
based and industrial. In our actual study, as mention, systems nowadays and the trend in this area.In our
we focus on the layered-IoT architectures and indus- current study, we choose to set the focus on the 3, 5,
trial architectures. Firstly, each layer address points and 7-Layers architecture models for several reasons.
separated before it is integrated and perform as a sys- Firstly, the 3-Layers model is the most abstract and
tem. This methodology helps manage the complexity straightforward model where the majority of research
of the system. IoT scenarios have a high level of com- in IoT architecture starts ((Khan et al., 2012) and (Wu
plexity because of the integration of various kinds of et al., 2010)). Secondly, the 7-Layers is considered
technologies, devices, objects, and services. the complete model, regarding the complexity of the
The primary studies designed their layered archi- IoT systems nowadays and the trend in this area. Fi-
tecture ranged from 3 to 7 layers which are composed nally, the 5-Layer model is a model between the other
by the main building blocks in the IoT platforms, two reference models (3 and 7-Layers model).
2.1.1 3-Layers Model • The first layer is where a variety of devices, sen-
sors, and controllers that enable their interconnec-
According to Ray (Ray, 2018) and Lin et al. (Lin tion are situated.
et al., 2017) the trivial architecture for IoT systems • The second layer is responsible for making all
is called 3-Layers as it has three layers: perception, connections and data transfers in the IoT system.
network, and application layers (Figure 2a). Thus, this layer specifies the communication pro-
On the Perception Layer are the things. The things tocols.
have sensors that are responsible for taking informa- • The third layer, the Edge/Fog Computing layer, is
tion, and actuators to interact with the environment. where the data analysis and data transformation,
The Networking Layer is in charge of connecting is performed (Tzafestas, 2018).
the thing with another things, network devices, and
• The fourth layer, the Data Accumulation, leads
servers.
with the storage data and guarantee that the data
The last layer on the architecture is the Applica- is moving correctly.
tion Layer that addresses the delivery of the services
for the final user, and it is on this layer that is the • The fifth is where the data is prepared to be ana-
clouds and servers. lyzed using the data mining techniques or the data
The IoT 3-Layers as an accepted structure but is implementation of machine learning (Tzafestas,
trivial modeling of the IoT ecosystem (Alshohoumi 2018). The last two layers are Application and
et al., 2019).One point important for 3-Layer archi- Collaboration & Process.
tecture is the fact that there is not a layer for Business. • The Application Layer is where the users can
use the information about the environment that is
2.1.2 5-Layers Model taken by the things.
• Finally, the seventh layer represents the actors that
Researchers proposed a new architecture to solve use the data to make a decision based on the data
the issues summarized in the 3-Layers architecture. extracted on the IoT ecosystem (Pisching et al.,
The 5-Layers (Figure 2b) is an extension of 3-Layers 2018).
with the introduction of Processing and Business Lay-
ers((Wu et al., 2010) and (Sethi and Sarangi, 2017)). 7-Layers

The characteristics and goals of the Perception Collaboration & Processes

and Application Layers are the same as the 3-Layers

architecture. The Transport Layer is responsible for Application

transferring the data from things to the Processing 5-Layers

Layer in both ways. Business Layer

Data Abstraction

The Processing Layer works as a middleware in

the 5-Layers architecture; the role of this layer is to Application Layer
Data Accumulation

store, analyze, and process the information of objects 3-Layers

received from the transport layer. With the Processing Application Layer Processing Layer Edge(Fog)Computing

Layer, various technologies can be applied, for exam-

ple, database, cloud computing, and big data process- Network Layer Network Layer Connectivity

ing modules.
Physical Devices &
Finally, the Business Layer is the layer that threat- Perception Layer Perception Layer Controllers

ens all IoT system management; it includes the ap- (a) (b) (c)
plication, business, and profit models and user pri- Figure 2: IoT Layers Architectures
vacy. In the 5-Layers, it is considered data storage
and processing, but neither security and privacy are
discussed.
3 RELATED WORK
2.1.3 7-Layers Model
A systematic review of existing IoT architectures was
The 7-Layers (Figure 2c) used in our current study performed by Alshohoumi et al. (Alshohoumi et al.,
was proposed in the Internet of Things World Forum 2019) on a set of 144 studies from 2008 to 2018. It
(IoTWF) (Pisching et al., 2018). Following the de- is a review of IoT architectures in terms of architec-
scription for each layer. ture classification (the number of layers), limitations
in each architecture, and considerations of different Based on the extracted data for each layer in each
aspects or features in each layer. Our study differs provider (input, output, activities, and main objective)
from the review perform by Alshohoumi et al. (Am- for the 7-Layers Model, we classified the layers in
mar et al., 2018) in the following terms: (i) we in- the industrial architecture according the layer in the
vestigate the architecture proposed by IoT providers, referential model. For example, for the architecture
(ii) we apply an approach to model and analyze the following by Microsoft, we compare each layer with
architecture following for IoT providers. the reference model using the data extraction for Mi-
Ammar et al. (Ammar et al., 2018) presents a sur- crosoft IoT white paper (Microsoft, 2018), and we
vey on security of IoT frameworks; a total of 8 frame- could classified each layer in the Microsoft proposal
works are considered. Their study had as a goal: clari- in most similar layer in the 7-Layers model.
fying the state of the art IoT platforms and identifying It is important to highlight there is some layer that
the trends of current designs of such platforms, pro- are not described neither in the documentation nor in
viding a high-level comparison between the different the figure represented the architecture. In this cases,
architectures of the various frameworks. Our study we decided to represent the layer using dotted line to
differs from their because, in our research, we model demonstrated that this layer does not have description
the IoT providers using our approach and before per- in the available documentation,
forming a comparison. In our approach for each layer in the IoT architec-
Pratap Singh et al. (Pratap Singh et al., 2020) ture in question, we extracted data to answer the fol-
presents an analysis of Layer-Specific, Domain- lowing set of questions using the information made
Specific, and Industry defined IoT architectures. The available for the IoT provider:
main contribution is a summarizing of state of the art • What is(are) the input(s) for the layer?
of IoT with a comparison of the industry-defined ar-
chitectures. The work by Pratap Singh et al. and our • What is(are) the output(s) for the layer?
study can be considered as complementary each of • What does the layer perform the activities?
them analyzes IoT architectures. Still, our study pro- • What is the principal objective of the layer?
poses an approach for modeling IoT architectures and
also compares seven industrial IoT architectures (In-
tel, Microsoft, Cisco, Google, and IBM).
5 MODELING SEVEN IOT
PROVIDERS USING OUR
4 APPROACH FOR MODELING APPROACH
AND ANALYZING OF IoT In this section, we apply the approach described in
ARCHITECTURES Section 4 for seven leading industrial actors: Intel,
Microsoft, Cisco, Google, IBM, Ericsson, and Ama-
In this section, we discuss our approach for model- zon.
ing and analyzing IoT architectures. The evaluation Our inclusion criteria for the IoT architectures
approach which we propose is built upon the referen- were selecting architectures that were studied at least
tial architecture described in the section 2. First, we twice by Ammar et al. (Ammar et al., 2018), Lueth
discuss the construction, and then we summarize the (Lueth, 2015) and Asemani et al. (Asemani et al.,
steps that composed our approach. 2019). We used these studies as a guideline to se-
We start analyzing the documentation that IoT lect the IoT architectures because of the following
providers make available publicly. These data are reasons: (i) these researchers perform a study simi-
available on websites or in white papers ((Intel, 2015), lar than our study, (ii) these studies analyzed deeply
(Microsoft, 2018), (Google, 2019), (IBM, 2019), the IoT architecture and (iii) these studies analyzed
(Cisco, 2014), (Amazon, 2016), (Ericsson, 2017)). the leading IoT companies around the world.
To analyze this information systematically, we
summarize essential aspects for each layer in the IoT 5.1 Intel Architecture
architecture by the provider. We read the documenta-
tion and conduct an extraction of some elements for Intel IoT Platform Reference Architecture is a system
each layer: input, output, activities performed, and architecture specification (SAS) to connect any prod-
the principal objective. It is important to note that ucts and services to the cloud (Intel, 2015). There
some aspects have different terminology or function- are two versions of this IoT Architecture. Firstly, ver-
alities in each layer for each provider. sion 1.0 is an architecture to connect the unconnected
things, and the version is the Intel SAS version 2.0 4) is not described neither in the documentation nor
that brings a specification about the integration of a in the layer architecture
variety of devices with intelligence and connectivity
integrated (Breivold, 2017).
In our study, we classified the Version 2.0 because
of it being a reference architecture that facilitated the
convergence of operational technology and informa-
tion technology. Additionally, it is a future-looking
reference architecture.
According to the (Intel, 2015), the three main
Figure 3: Intel End-to-End IoT Solution form Things to
components are: things, network and cloud (Figure Network to Cloud. (Intel, 2015)
3). The Things component is not represent in the Lay-
ered Architecture but as we notice in the documenta-
tion available, Intel defined a end-to-end solution for 5.2 Microsoft Architecture
connecting nearly any type of device to the cloud.
In Figure 4 more specifically in the figure more The design of the Microsoft IoT architecture is based
in the right and dotted, we have the architecture pro- in three layers: Thing, Insight, and Actions. There
poses by Intel (Intel, 2015). The first group of layers are a set of subsystems in all layers, and it is essen-
(in dark blue) have layers that are the major run-time tial to highlight that Edge Devices, Data Transforma-
layers. The Communication and Connectivity layer is tion, Machine Learning, and User Management are
situated on the bottom of the architecture, and are re- optional subsystems (Microsoft, 2019).
sponsible for enabling multi-protocol data communi- The architecture proposes by Microsoft (Figure
cation between not only devices and the edge but also 1b) the first group of subsystems are in the layer
between endpoint devices/gateways, the network, and “Things” and are IoT Devices, IoT Edge Devices, and
the data center (Intel, 2015). Device Provisioning.
The second layer is the Data layer with Analyt- The IoT devices are devices that need to register
ics, whose role is providing customer value. It is in a way secure with the cloud to send and receive
achieved using valuable insights generated by data data. The second group is on the layer “Insights” and
analytics and improved closed-loop control systems. consists of Cloud Gateway, Data Transformation, UI
On the Intel IoT reference architecture, this need and Reporting Tools, Stream Processing, Warm Path
is addressed by allowing analytics to be distributed Store, Cold Path Store.
across the cloud, gateways, and smart endpoint de- In the case of the devices or field gateways that
vices (Breivold, 2017). are not able to use the standard protocols used by IoT
The Management layer is the next layer in which Hub, adaptation is necessary, and Azure IoT protocol
the primordial role is for realizing automated discov- gateway can be used (Microsoft, 2019). After that, we
ery and provisioning of endpoint devices. Intel rec- have in the Microsoft IoT architecture the Stream Pro-
ommends Device Cloud product to perform the man- cessing that is responsible for consuming the massive
ageability functions (Breivold, 2017). And to com- streams of data records and evaluate rules for those
plete the dark blue group, the Control layer that pro- streams.The third layer is the “Action” layer. On this
vides a way to separate the management layer into a layer, we have the Business integration, which essen-
management plane and control plane, with policy and tial role is performing actions based on the teleme-
control objects and APIs. try data during stream processing. Also, in the “Ac-
Now, the second group is composed of user lay- tion” layer, we have the User Management subsystem
ers. The first is the Application layer that is used by whose goal is restricting the user or user groups ac-
the Business Layer to access the other layers in the tion on the devices. And finally, we have the Machine
Intel IoT architecture. Besides these layers, there is a Learning subsystem that is responsible for perform-
vertical security layer. The role of the security layer ing predict algorithms using the telemetry data. This
is to address protection and security across all tiers. prediction can be applied in predictive maintenance,
We could summarize some aspect about Intel ar- for example.
chitecture, applying our approach. The first aspect We apply our approach and could summarize sim-
important in the Intel architecture is that there is a ilarities and divergences between the layers in the Mi-
Business Layer, what shows that for Intel is impor- crosoft architecture and the layers in the 7-Layers. We
tant to have clearly the IoT economic impact. And, in could find similarities between the layers IoT Edge
Intel architecture, the Data Abstraction layer (Figure Devices and Cloud Gateway and the Transport layer.
 Intel Architecture 

7 Business Layer

Application Layer
6
Control Layer

5 Not described

6
4 Management Layer

4
3 Data Layer (Analytics)

3
Communications and
2
Connectivity Layer

2
1 Things

Figure 4: Analyze of Intel IoT Architecture.

The Cloud Gateway and Data Transformation words, the data is moving at the rate and organization
modules in the architecture follows by Microsoft have defined by the devices that generated the data in level
similarities with the layer Edge(Fog)computing in the 1. Data Accumulation level will make network data
referential model 7-Layers. The Data Accumulation turn into data usable for the application. The idea is
is represented by the Warm Path Store and Cold Path to convert data-in-motion to data-at-rest, converting
Store. For the Data Abstraction in the referential data network packets to relational database tables.
model, we have the Stream Processing in the Mi- After the data accumulation is necessary to ren-
crosoft architecture. der the data and store this information in a way to
For the Application layer in the 7-Layers model, facilitate the development of application more simple
we could find similarities with the modules UI and and with higher performance, this is performing by
Report Tools, User Management and Machine Learn- the level 5.Level 6 is the application level, where all
ing. Finally, the activities that composed the Business the data that is generated is interpreted using various
layer are performed by the module Business Integra- types of an application. The collaboration & process
tion. level execute the applications with the specific needs.
With the apps (level 6), people have access to the right
5.3 Cisco Architecture data at the right time so that they can do the right thing
(Cisco, 2014).
We applying our approach and find just similari-
Cisco proposal(Figure 6)is a multi-level reference that
ties between the layers in the Cisco architecture and
is composed of seven layers in the dotted area shows
the layers in the 7-Layers. For us, the analysis of
(Cisco, 2014).
Cisco was proof that our approach is a valuable tool
In level 1 is the Things on the Internet of Things. to be used to model and analyze IoT architectures.
As an example, we can have endpoint devices that
may send and receive information. Level 2 is focused
on communication and connectivity. The connectivity 5.4 Google Architecture
includes the data transmission between devices and
the network, across networks and between the net- Google Cloud Platform (GCP) is a complete set of
work and low-level information processing. The ac- tools to connect, process, store, and analyze data both
tivities in level 3 are focused on reorganizing the data at the edge and in the cloud (Google, 2019). To GCP,
that is produced by the things in the level 1 into in- the IoT solution consists of three essential compo-
formation that is convenient for level 4 (storage and nents, the device, gateway, and cloud. The element
higher processing). The priority of level 4 is guar- that is directly related to the world is the device. For
anteeing that the data is moving precisely. In other this reference architecture, the device can be hardware
 Microsoft Architecture

5 6 7 Business Integration 
6
2
UI and Report Tools,
6 Machine Learning and User
3 4 Management
2 3
7
5  Stream Processing

1
Warm Path Store and
4 Cold Path Store
4
6

3 Cloud Gateway  and Data

Transformation

IoT Edge Devices 

2
and Cloud Gateway

1 IoT Devices

Figure 5: Analyze of Microsoft IoT Architecture

Cisco Architecture 
7 Collaboration and Process

6 Application

7
5 Data Abstraction
6

5 4 Data Accumulation

4
3 Edge(Fog) Computing
3

2
2 Connectivity

1
1 Physical Devices and
Controllers

Figure 6: Analyze of Cisco IoT Architecture.

and software and might be directly or indirectly con- are three modules in the Data Analytics in the Cloud
nected to the Internet. Besides that, devices can com- that can be classified as Data Accumulation in the
municate with each other via a network. 7-Layer model Although Google uses only four ele-
ments to present the IoT architecture following in its
The gateway is responsible for guaranteeing the
solutions, it was necessary to ”open” the modules to
devices that are not directly connected to the Internet
understand deeply what the activities performed and
to reach cloud services. Another aspect of the gate-
how we could classify using our approach.
way is that it processes data on behalf of a group or
cluster of devices. The devices collect data; the gate-
way sends this data to Cloud Platform, the third part
5.5 IBM Architecture
of the architecture. On the cloud platform, the data
The IBM (Figure 8) proposal for IoT architecture is
are processed and combined with other data that were
composed by the following layers User Layers, Prox-
sent for different devices.
imity Network, Public Network, Provider Cloud, and
We analyzed the similarities and divergence be- Enterprise Network. The main objective of IBM IoT
tween Google architecture and our referential model architecture is to perform a connection to IoT de-
(Figure 7). In our analysis, we concluded that there is vices and quickly build scalable apps and visualiza-
a high level of linkage between the internal modules tion dashboards to gain insights from IoT data. IBM
in the Data Analytics in the Cloud. For example, there IoT architecture uses IBM Cloud IoT, data, and AI
 Edge gateway Data Analytics in the Cloud Data Usage

Update device
config 6
IoT Devices Update
device 4
config &
deploy ML 3 7 Insights
model and
container 4
Control
Data Studio and Cloud
Data Data 6 DataLab

4
5 Cloud Machine Learning
Control Data Update device Training
5
config 7
Cloud(Pub/Sub,
4
Bigtable,DataFlow)
Serving

3 Cloud IoT Core, Functions

2 Edge Gateway
Direct cloud
Data connectivity
through
Cloud IoT 1 IoT Devices
Core
Security libraries
Connectivity
libraries
Cloud IoT Control
Providiong agent
RTOS Kernel
HW Driver

Powered by Cloud
IoT Device SDK

Figure 7: Analyze of Google IoT Architecture.

services to achieve the main goal (IBM, 2019). performing as one layer in the referential model.
There are five layers called the User Layer, Prox- Edge Service, IoT Transformation & Connectivity,
imity Network, Provider Cloud, and Enterprise Net- and Transformation & Connectivity compos the third
work. We could find similarities between the layers layer in the IBM architecture as is comparable with
in the IBM architecture and the 7-Layers model with the Edge(Fog) Computing in the 7-Layers model.
the application of our approach. But it is essential to The modules Data Store and Enterprise Data num-
highlight that we made the analysis using the inter- bering perform activities similarities with the Data
nal modules insight the layers: User Layer, Proximity Accumulation. Application logic (IBM architecture)
Network, Provider Cloud, and Enterprise Network. is the Data Abstraction(7-Layers model).
The sensors (water leak detection, water flow, All device modules (registry, management, and
temperature, etc.) and actuators (automatic water identity service), API Management, User Directory,
shutoff valves) are in the Proximity Network; this Process Management, Analytics, and Visualization
layer can be at home for example, if we talk about compos the sixth layer in the IBM architecture and
IoT platforms for Smart Homes. They are attached to is comparable with the Application Layer in the 7-
the device maker’s cloud service. Layers model.
In this scenario, we can see an IoT application, the Finally, IoT User and Enterprise Application are
Smart homes. The connected devices let insurance the seventh layer in the IBM architecture and are com-
companies improve service for their policyholders. parable with the Collaboration and Processes in the
Besides that, with data collected is possible to provide 7-Layers.
insight about risks that can happen in the home. For
example, leak-detection sensors can monitor for water 5.6 Ericsson Architecture
leaks. With a correct analyze, valves can be trigged,
and the IoT platform can help protect the home from Ericsson’s IoT architecture proposal (Figure 9) is a se-
resulting damage. curity solution that provides continuous monitoring of
The Device performs activities comparable with threats, vulnerabilities, risks, and compliance, along
the Perception Layer. The IoT Gateway is the Trans- with automated remediation (Ericsson, 2017).
port Layer in the IBM architecture. We analyze the architecture proposal by Ericsson
In the layer numbering with three, we have more using our methodology. We classified the IoT devices
than one module in the industrial architecture (IBM) as the Perception layer, the IoT Gateway, Access, and
 IBM Architecture 

7 IoT User, Entereprise Application

6
Application, Device (Registry, Management Identity
Service), API Management, User Directory, Process
6 6 Management, Analytics, Visualization
7 6 6
1

3 3 4 5 Application Logic
6
2
3 Data Store, Enterprise Data
4
5 7

6 3 Edge  Service, (IoT) Transformation & Connectivity

6 6 6 4

2 IoT Gateway

1 Device

Figure 8: Analyze of IBM IoT Architecture.

network connectivity as a Network layer. In the layer model. As well as in the Ericsson architecture (Sub-
IoT app, platform & cloud, we could find similarities Section 5.6), there is no description of a Business
between the activities performed in the module Cloud Layer in the Amazon architecture.
Infrastructure with Edge(Fog)Computing, and Appli-
cation with Data Abstraction, we do not find the de-
scription about the Data Accumulation. IoT User as
Application Layer.We do not find represent the Busi-
6 ANALYZING SEVEN IOT
ness Layer. PROVIDERS USING OUR
The Business Layer and Data Accumulation in the APPROACH
Ericsson architecture is not represented but all the
other layers have similarities with the layer in the 7- This section presents a comparative analysis of seven
Layers model. industrial architecture modeled in Layer-Model using
our approach (Figure 11).
5.7 Amazon Architecture Although the fact the seven industrial architec-
tures have terminology different, it was possible to
The main objective of the AWS IoT is providing se- compare the layers providers to the layer-model us-
cure, bi-directional communication between Internet- ing the methodology that we proposed. We noticed as
connected devices such as sensors, actuators, embed- far as that the abstraction level concerning the physi-
ded micro-controllers, or smart appliances, and the cal layer increases, the terminology became more di-
AWS Cloud (Amazon, 2016). verse.
The elements that compose the architecture are We could classified all the layers in the seven ar-
Alexa voice service integration for AWS IoT, custom chitectures studied into seven layers. In other words,
authentication service, device gateway, device provi- we use 7-layers model to identify the layers in the in-
sioning service, device shadow, device shadow ser- dustrial architecture that can be considered similar.
vice, group registry, jobs service, message broker, For the network layer in the 7-Layers model, we
registry, Rules engine and security, and integrity ser- notice that five of the seven industrial have used the
vice (Amazon, 2016). term Gateway or similarities to express the layer re-
We analyze the architecture proposal by Ama- sponsible for performing the communication with and
zon(Figure 10) using our approach and classified the between the perception layer. It is important to clar-
component Devices as a Perception Layer. The Ama- ify the gateway has a specific function in networking,
zon documentation describes the Gateway compo- it is also used to describe a class of device that pro-
nent, but it does not show in the architecture. The cesses data on behalf of a group or cluster of devices
core of the AWS solution (Guth et al., 2016) is com- (Google, 2019).
posed of the components numbering by 3, and they Five industrial architectures defined, in the docu-
have similarities with the Processing in the 7-Layers mentation, the Business layer, and three of them use
 Ericsson Architecture 

7 Not described
1 2 2 6

6 IoT user (Application)

5 IoT app (Application)

4 Not described

3 Cloud Infrastrucutre

IoT Gateway, Access 

3 2 and network connectivity

1 IoT Device

Figure 9: Analyze of Ericsson Architecture

the term Business or similar to denoted the last level 7 THREATS TO VALIDITY
of them architectures. We conclude that for the ma-
jority of the companies is important to represent the During our study, we have counted some threats that
economic impact of the IoT solutions. Another aspect we need to address. We work carefully to mitigate
in the last layer for the industrial architectures and be- eventual issues that could compromise the validity of
come a recommendation for Ericsson and Amazon is our results or conclusions. In this section we highlight
making explicit the Business layer in them architec- some of those threats and what mechanism that we
tures. applied to address it.
Our study could show in a comparative approach First, the main limitation of this work is the anal-
the fact that is known in the IoT community: there ysis of the IoT architecture by the providers was per-
is no single consensus on IoT architecture. The com- forming using only the information available in sites
plexity of IoT architecture is one of the reasons for and white paper. However, we opted to perform
many proposal architectures. With decreasing com- the analysis only with the data made available for
plexity, using a referential model, we could evaluate the companies because it is this information that our
different architectures. audience also has to analyze and compare the IoT
providers. Besides, we could extract the data, pro-
We learned during the IoT architecture study some poses an approach, and compare the architectures.
major lessons: The second threat to validity is the bias created by
the fact that we used one type of referential model,
• The use of layers to represent architecture is a way and we chose three models. However, as this, our
more clear to understand the architectures. mapping is preliminary work; consequently, we de-
cide to start the study only with these models and af-
• There are some layers in the industrial IoT archi- ter as a future work performing the same analysis with
tectures analyzed that perform the same activities other models.
but with different nomenclature.

• There are modules in industrial architecture that 8 CONCLUSION

perform activities inherent to more than one layer
in the layer-model architecture. IoT market is continually growing; as a consequence,
the number of IoT architectures has increased too.
• It is necessary a standard to provide a better com- There is no unique or consensus about the IoT archi-
prehension of the issues and threats in IoT. tecture; each architecture has the approach and the in-
 Amazon Architecture 

7 Not described
3
3
1
5
6 IoT Applications

3 5

5 Amazon (DynamoDB, Kinesis, Lambda, SNS, SQS)

5

4 Amazon S3
4

6
3
5 3 Gateway, Message broker, Device Shadows, Rules Engine 

5
2 Device Gateway

1 Devices

Figure 10: Analyze of Amazon Architecture: the Business Layer is not clearly represented in the architecture documenta-
tion((Amazon, 2016)) but all the other layers have similarities with the 5-Layers model.
Microsoft Architecture Cisco Architecture IBM Architecture Google Architecture

7 Business Integration  7 Collaboration and Process 7 IoT User, Enterprise Application 7 Insights

Application, Device (Registry, Management, Identity

6 UI and Report Tools,Machine 6 6 6
Learning  and User Management
Application Service), API and ProcessManagement, User Directory, Data Studio and Cloud DataLab
Analytics and Visualization

5  Stream Processing  5 Data Abstraction 5 Applicaiton Logic 5 Cloud Machine Learning

Warm Path Store and 

4 4 Data Accumulation 4 Data Store and Enterprise Data 4 Cloud(Pub/Sub, Bigtable,DataFlow)
Cold Path Store

Cloud Gateway and 3 Edge Service, (IoT) Transformatiomn &

3 Edge(Fog) Computing 3 Cloud IoT Core, Functions
 Data Transformation 3 Connectivity

2 Cloud Gateway 2 Connectivity 2 IoT Gateway 2 Edge Gateway

1 IoT Devices and IoT Edge Devices 1 Physical Devices and Controllers 1 Device 1 IoT Devices

Ericsson Architecture Amazon Architecture

Intel Architecture
7 Business Layer Not described 7 Not described
7

Application Layer
6 6 IoT user (Application) 6 IoT Applications
Control Layer

5 Not described 5 Amazon (DynamoDB, Kinesis, Lambda,

IoT app (Application) 5
SNS, SQS)

4 Management Layer 4 Not described 4 Amazon S3

3 Data Layer (Analytics) Gateway, Message broker, Device Shadows,

3 Cloud Infrastrucutre
3 Rules Engine 

2 Communications and Connectivity Layer 2 IoT Gateway, Access  2 Device Gateway

and network connectivity

1 Things 1 IoT Device 1 Devices

Figure 11: Industrial providers classified as a Layers-Model

terpretation of IoT for the provider. In this study, we summarize information about
seven IoT industrial architectures. And, we provide Ammar, M., Russello, G., and Crispo, B. (2018). Inter-
an approach that makes possible the modeling of IoT net of things: A survey on the security of iot frame-
industrial architectures based on Layers-Model, as works. Journal of Information Security and Applica-
well as the comparison between IoT architectures. tions, 38:8 – 27.
Asemani, M., Abdollahei, F., and Jabbari, F. (2019). Un-
With our analysis, we could concluded that even
derstanding iot platforms : Towards a comprehen-
though the industrial architectures have represented sive definition and main characteristic description. In
IoT solutions in a different way with a different termi- 2019 5th International Conference on Web Research
nology, architectures perform the same activities. Be- (ICWR), pages 172–177.
sides that, it was clear that some architectures follow- Breivold, H. P. (2017). A survey and analysis of reference
ing the seven-layers model, for example, Microsoft architectures for the internet-of-things. ICSEA 2017,
and IBM) but others are used the model differently, page 143.
for example, Intel and Google. Cisco (2014). The Internet of Things reference model.
We could conclude not only the need for a stan- http://cdn.iotwf.com/resources/71/IoT Reference
Model White Paper June 4 2014.pdf.[Online; ac-
dard for IoT architecture and but also the advantages cessed 06-January-2020].
to represent the architectures using layers. With the Ericsson (2017). End-to-end security manage-
layers, it was easier to compare and analyze better the ment for the IoT. https://www.ericsson.
industrial IoT architectures. com/4ab42a/assets/local/reports-papers/
We aim that our study benefit researchers that per- ericsson-technology-review/docs/2017/
forming analysis of IoT architectures and practition- managing-iot-security-end-to-end-etr-10-2017.pdf.
ers that need to choose IoT providers. Future work in- [Online; accessed 20-December-2019].
cludes (i) a quantitative analysis of IoT architectures Google (2019). Overview of internet of things. https://
cloud.google.com/solutions/iot-overview.[Online; ac-
studied, (ii) a systematic literature review of Internet
cessed 11-December-2019].
of Things architectures, and (iii) systematic analysis
Gubbi, J., Buyya, R., Marusic, S., and Palaniswami, M.
about the products that are available by the providers. (2013). Internet of things (iot): A vision, architec-
tural elements, and future directions. Future Gen-
eration Computer Systems, 29(7):1645 – 1660. In-
cluding Special sections: Cyber-enabled Distributed
ACKNOWLEDGEMENTS Computing for Ubiquitous Cloud and Network Ser-
vices & Cloud Computing and Scientific Applications
To the members of the SmArtSE Research Team, Big Data, Scalable Analytics, and Beyond.
Synchromedia Laboratory and Computer Research Guth, J., Breitenbcher, U., Falkenthal, M., Leymann, F., and
Institute of Montral (CRIM) for their support Reinfurt, L. (2016). Comparison of iot platform archi-
and knowledge sharing. This work was financed tectures: A field study based on a reference architec-
ture. In 2016 Cloudification of the Internet of Things
by the Canadian program MITACS and LabVI
(CIoT), pages 1–6.
(http://quartierinnovationmontreal.com/en/open-sky-
IBM (2019). Internet of things for insights from con-
laboratory-smart-life). nected devices. https://www.ibm.com/cloud/garage/
architectures/iotArchitecture.[Online; accessed 06-
December-2019].
Intel (2015). The Intel IoT Platform. https://www.intel.com/
REFERENCES content/www/us/en/internet-of-things/white-papers/
iot-platform-reference-architecture-paper.
Ahmed, B. S., Bures, M., Frajtak, K., and Cerny, T. (2019). html.[Online; accessed 17-December-2019].
Aspects of quality in internet of things (iot) solutions: Khan, R., Khan, S. U., Zaheer, R., and Khan, S. (2012). Fu-
A systematic mapping study. IEEE Access, 7:13758– ture internet: The internet of things architecture, pos-
13780. sible applications and key challenges. In 2012 10th
Al-Fuqaha, A., Guizani, M., Mohammadi, M., Aledhari, International Conference on Frontiers of Information
M., and Ayyash, M. (2015). Internet of things: A Technology, pages 257–260.
survey on enabling technologies, protocols, and ap- Lin, J., Yu, W., Zhang, N., Yang, X., Zhang, H., and Zhao,
plications. IEEE Communications Surveys Tutorials, W. (2017). A survey on internet of things: Archi-
17(4):2347–2376. tecture, enabling technologies, security and privacy,
Alshohoumi, F., Sarrab, M., Al-Hamdani, A., and Al-Abri, and applications. IEEE Internet of Things Journal,
D. (2019). Systematic review of existing iot architec- 4(5):1125–1142.
tures security and privacy issues and concerns. Inter- Lueth, K. L. (2015). The top 20 Internet of
national Journal of Advanced Computer Science and Things companies right now: Intel over-
Applications, 10. takes Google. https://iot-analytics.com/
Amazon (2016). AWS IoT Documentation. https://docs.aws. 20-internet-of-things-companies/.[Online;accessed
amazon.com/iot/.[Online; accessed 04-March-2020]. 18-February-2020].
Microsoft (2018). Azure iot reference architecture.
https://aka.ms/iotrefarchitecture.[Online;accessed 16-
December-2019].
Microsoft (2019). Supoort additional protocols for
iot hub. https://docs.microsoft.com/en-us/azure/
iot-hub/iot-hub-protocol-gateway.[Online; accessed
16-December-2019].
Muccini, H. and T. Moghaddam, M. (2018). IoT Architec-
tural Styles, pages 68–85. 12th European Conference
on Software Architecture, ECSA 2018.
Pan, J. and McElhannon, J. (2018). Future edge cloud and
edge computing for internet of things applications.
IEEE Internet of Things Journal, 5(1):439–449.
Pisching, M., Pessoa, M., Junqueira, F., Santos Filho, D.,
and Miyagi, P. (2018). An architecture based on rami
4.0 to discover equipment to process operations re-
quired by products. Computers & Industrial Engi-
neering.
Pratap Singh, S., Kumar, V., Kumar Singh, A., and Singh,
S. (2020). A survey on internet of things (iot): Layer
specific vs. domain specific architecture. In Smys,
S., Senjyu, T., and Lafata, P., editors, Second In-
ternational Conference on Computer Networks and
Communication Technologies, pages 333–341, Cham.
Springer International Publishing.
Ray, P. (2018). A survey on internet of things architectures.
Journal of King Saud University - Computer and In-
formation Sciences, 30(3):291 – 319.
Sethi, P. and Sarangi, S. (2017). Internet of things: Archi-
tectures, protocols, and applications. Journal of Elec-
trical and Computer Engineering, 2017:1–25.
Tzafestas, S. G. (2018). The internet of things: A concep-
tual guided tour. European Journal of Advances in
Engineering and Technology, 5(10):745–767.
Wu, M., Lu, T.-J., Ling, F.-Y., Sun, J., and Du, H.-Y.
(2010). Research on the architecture of internet of
things. In 2010 3rd International Conference on Ad-
vanced Computer Theory and Engineering (ICACTE),
volume 5, pages V5–484. IEEE.