ABSTRACT: network. IoT environments are described by a

ABSTRACT: The Internet of Things (IoT) is
a epitome in which smart objects actively communicate with diverse physical
devices without the need of human interaction that are in a network. IoT environments
are described by a grade of miscellaneous including devices with diverse
capabilities, Operational and network protocols.To handle such heterogeneity
some platforms have proposed to elevate interoperability by eliminating the
explicitness of such devices. In this context the referral architecture is used
for outlining the inaugural set of building blocks for the IoT environment and
provides a solid establishment. This paper targets on reviewing the distinctive
reference architectures and provide well integrated and context aware
intelligent services for IoT.

 

KEYWORDS: Internet-of-things,
IoT, interoperability, reference architecture.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

 

                                                                                                                                     
i.     
INTRODUCTION

 

The
Internet of Things has become the recent trend since it’s the smartest way of
connecting the physical things or devices to the digital world 1. As the
count of devices and things connected to internet has been increased, by 2020
the count of internet connected things will outnumber humans by 4-to -12.IoT
are basically portrayed by a high grade of hardware and software
miscelaneousnesss .IoT platform proposed mainly concentrate on absolute
integration of miscellaneous devices and accoutre the high level models for
developing applications 3, 9. The lack of standard in the IoT context makes
the solution to adapt different programming models that are not congenial with
each other and thus arises scalability issues.. To overcome those above
mentioned limitations reference architecture is required for outlining the
initial set of building blocks for the IoT Environment and provide a solid
establishment for the IoT.

 

In
this paper a survey is done on some existing reference architectures for the
Internet of things and analysed them for various kind of perspectives such as
platform, security,  edge computing,authentication
and real-time analysis

 

 

Reference
architecture can be said as abstract architecture encompassing knowledge and
skills in a accustomed application domain thus making ease and walk through
development, standardization, interoperability, and progression of software
system in such a domain 4. Establishing of reference architecture is in need
as they depict both integral building elements and design choices for dealing
with functional and non-functional essentials in IoT environment. Reference
architecture allow the building systems 
be able to properly satisfy the existing conditions by considering  the set of building blocks provided by such
architectures. Maintaining interoperability solutions is the main concern in
IoT environment. Interoperability is attained by building architectures based
upon reference architecture. The IoT architecture portrayed in  Figure 1 5 is a 4 tier architecture (IoT).

 

 

 

 

 

 

Figure
1 IoT reference Architecture

 

 

 

The 4 tier
architecture is described below. 

 

Stage -1

 

Sensors/actuators

Sensor is a
device that detects and responds to some type of input data from the physical
environment .For example temperature sensor sense the temperature and sends the
data. Actuators can also interpose to change the physical conditions that
generate the data. For example an actuator can shut down a power supply, adjust
an air valve.

 

 

Stage-2

                      The Internet Gateway.

 

The raw
data(analog) from the sensors is summated and transformed to digital stream for
further course of processing . Data summation and conversions are done by data
acquistion systems .The  digitized and
aggregated data  is received by the
gateway  that has to be sent to next
stage for further. Gateway often sits in closeness to the sensors and
actuators.

 

Edge IT

The digitized
and aggregated data from the 2nd stage is forwarded to 3rd stage for the further
processing .This is the stage before the data enters into the data
center.Pre-processing and analysing of data is done in this stage .Because to
decrease the load on  IT core
infrastructure and to save the network bandwidth. Machine learning can be used
at the edge to scan for abnormalities. That identifies impending maintenance
problems that need immediate attention , then by using visualization   technology the information can be presented
using dash boards, graphs or maps.

 

Stage-4

 

                 
The Datacentre and Cloud

Data that needs
to be more in depth processed and feedback doesn’t have to be immediate gets
forwarded to data centre or cloud based systems. Here high processing  IT systems 
are used for  analysing ,
managing  and securely storing  the data. Longer time is required for getting
the results in 4th stage but more in-depth analysis is done in this stage. In
this stage data from sensors can be appended with the data from the other
sources to give deeper insights.

 

 

                                                                                                                           
II.           
AWS IOT REFERNCE
ARCHITECTURE

 

 

AWS is an amazon
web services platform that collects the Telemetry data from the internet
connected things and analyses it and also provides secure bidirectional
communication between internet catenated devices and the cloud. Its
architecture is displayed in Figure 26. AWS 
consists of these components 1) Device gateway 2) Message broker 3) Rules
engine4) Security and Identity service5) Thing shadow 6) Thing
Shadows service. It provides SDK for connection of  hardware devices or mobile applications and it
let the devices to connect , authenticate and exchange messages using  the protocols (MQTT,HTTP,WEB Socket).It has a
device gateway using which messages can be exchanged using
publisher-subscription model. It has device shadows mode which consists of the
device latest state and it makes easier to build applications and those
interact with devices by providing available REST
API’s.The rules Engine helps to make IoT applications that collects, process,
and analyse and it acts on the data produced by the connected devices at global
scale without managing any infrastructure and it is also used for transforming
and routing  the IoT traffic to a precise
location or endpoint such as S3 function or lambda function.  AWS is cloud service provider multiple data
processing services are already integrated

 

 

                                                                   Figure
2 AWS IoT Reference Architecture

 

                                                                                                                  
III.           
Microsoft Azure IoT Reference
Architecture

The Azure IoT
is built on Microsoft azure platform to connect ,store ,analyse  and operationalize device data to give
business insights11 .It is a 3 level architecture which consists of 1) Device
connectivity 2) Data processing and Analytics 3) Presentation and business
connectivity . IoT hub provides a device registry that maintains a list of
devices and provides access to the device specific queue   Its architecture is displayed in figure
37,8.The principal point of this reference architecture is to provide
software and hardware heterogeneity for different scenarios and provide
security15 .Device connectivity provides the different devices to be
connected to the IoT cloud using protocols such as AMQP, HTTP and MQTT. Data
processing and analytics process the telemetry data that is obtained from
devices and analyses it. Presentation and business connectivity gives the end
clients to see and break down the information from the Devices. There is
additionally great support for IoT checking and diagnostics .Event hub
alongside the Azure stream examination is utilized for real-time analysis                                                                           

 

Figure
3 Azure IoT Reference Architecture

 

 

 

 

                                                                                                                   
IV.           
Google IoT
Reference Architecture

 

Google cloud Internet of things
provides a fully handled service for secure connection and managing of IoT
devices. It is three tier architecture 10 that consists of 1) Devices 2)
Standard devices (gateway), 3) Cloud Platform. The telemetry data enters through
cloud publisher/subscriber (pub/sub)12. Cloud pub/sub supports large scale
message ingestion over HTTP (or) gRPC. Ingested data is maneuvered by the
pipeline. Pipeline provides the processing tasks a) Transforming data b) Data
aggregation and computing c) Enriching data and d) Moving data. Cloud dataflow
renders service for data to be processed in various ways. And it is also
devised to assimilate seamlessly with other cloud platforms. Data is stored
along Enterprise analytics and logs using big query . Cloud Data lab is a tool
for exploring, transforming, analysing and Visualisation of data.  State of devices are automatically
synchronised with help of NoSQL cloud hosted database. It has built in
operational logging and monitoring assistance.

 

 

 

Figure
4 Google IoT Reference Architecture

 

 

                                                                                                                       
V.           
WSO2 IoT
Reference Architecture

 

WSO2
is open-source technology that provides the design for the IoT. That aims to
support amalgamations between systems and devices11,14. The architecture is
portrayed in Figure 513. The WSO2 architecture has five layers –

        1. Devices – Device layer in which
every device has a distinctive identity and it is either straightly or
indirectly linked to internet.

        2. 
Relevant transports- Communication layer which provides the connectivity
of the devices using different protocols.

        3. Aggregation /bus layer –
Aggregation/bus layer which aggregates communications from multiple devices,
brokers communication to a particular device , and transforms between various
protocols

        4. Event processing & analytics –
event processing and analytics layer,which processes and acts upon events from
the bus, and perform data storage .

        5. Client /external communications –
which empower users to communicate and interact with devices and obtain
views into analytics and event processing.

 

Besides the
vertical layers it has also two cross cutting layers –

a) Device
manager-Which communicates with remotely manages devices, and maintain the
identities of the devices.

b) “Identity and
access management” -It’s for access control.

 

 

Figure
5 WSO2 Reference Architecture

 

                                                                                                                                       
VI.           
Analysis and
Discussion

 

The referral architectures
depicted in the section two has resemblance in their technical aspects,
architectural principles and they differ in their technology approaches and
implementation. Therefore we categorized some particular characteristics that
have resembling concerns to describe the same or related facets of the reference
architecture. The facets in which the reference architectures are compared are

·        
Platform
– which covers like what are the protocols being used, Authentication, SDK,
communication pattern the architecture following and Backup.

·        
Real
time analytics

·        
Edge
computing

·        
Security

·        
Device
level authentication

Table 1 outlines the analogizing
of the referenced architectures.

 

 

                                                                                                                                                               
VII.           
Conclusion

 

IoT
platforms are growing their influence in daily life as the devices connecting
to the internet has been increasing and the integration of heterogeneous
hardware to the cloud is also increasing,so a referal architecture has to be
developed. In this paper a survey was done on some existing reference
architectures of IoT , and comparison has been made on those architectures
based on some facets of   the
architecture.

                For security WSO2 is not
providing TLS level of security compared to others and for the device SDK also
it is providing only in java .WSO2 provides extensive support for protocol and
authentication. Only AWS and WSO2 are providing edge computing in the
architecture while remaining are providing along with the SDK’s.

 

 

                                                                                                                                                        
VIII.           
References

 

1
Khan,
R. et al., “Future Internet: The Internet of Things Architecture, Possible
Applications and Key Challenges,” in Proceedings of the 10thInternational
Conference on Frontiers of Information Technology. IEEE,Dec.2012

2https://www.gartner.com/imagesrv/books/iot/iotEbook_digital.pdf

 

3 S. Bandyopadhyay, M. Sengupta, S. Maiti, and S.
Dutta. Role of middleware for Internet of Things: A study. International
Journal of Computer Science & Engineering Survey, 2(3):94-105, Aug. 2011.

4 E. Y. Nakagawa, F. Oquendo, and
J. C. Maldonado.Reference architectures. In M. C. Oussalah, editor, Software
Architecture 1, pages 55{82. ISTE Ltd /John Wiley & Sons, Inc., United
Kingdom, 2014.

5https://techbeacon.com/4-stages-iot-architecture

 

6https://aws.amazon.com/documentation/iot/

 

7
https://docs.microsoft.com/en-in/azure/iot-suite/iot-suite-what-is-azure-iot

8 Azure IoT Reference Architecture.
© 2016 Microsoft
Corporation. All rights reserved.

 

9 P.
F. Pires, E. Cavalcante, T. Barros, F. C. Delicato,T. Batista, and B. Costa. A
platform for integrating physical devices in the Internet of Things. In Proceedings
Of the 12th IEEE Int. Conf. on Embedded and Ubiquitous Computing,

pages
234-241, USA, 2014. IEEE.

10https://cloud.google.com/solutions/iot-overview

 

11 Hongyu Pei Breivold , “Survey and Analysis of Reference Architectures for
the Internet-of-thing”in proceedings of ICSEA 2017 : The
Twelfth International Conference on Software Engineering Advances.

 

 12 https://cloud.google.com/solutions/iot-overview#telemetry

 

13 Paul Fremantle “Whitepaper
of WSO2- A REFERENCE ARCHITECTURE FOR THE INTERNET OF THINGS”  VERSION 0.9.0(OCTOBER 20,2015).

 

14http://docs.wso2.com/display/IoTS310/Architecture

 

15 Jasmin
Guth, Uwe Breitenb¨ucher, Michael Falkenthal, Frank Leymann, and Lukas Reinfurt
 ”Comparison
of IoT  Platform Architectures: A Field Study
based on a Reference Architecture”. ©
2016 IEEE Computer Society

 

 

 

 

 

 

 

 

 

Table
1 Comparison of Reference Architectures of IoT

 

parameters
 

Azure
 

AWS
 

google
 

WSO2

1
Platform

Protocols

MQTT,AMQP and HTTP

MQTT,HTTP  1.1

Weave,MQTT,HTTP

MQTT,HTTP,
AMQP

 

SDK

.NET,UWP,Java,C
,NodeJS,Python

C,NodeJS,Java
,Python,IOS
 

Java,Python,NodeJS,Ruby
Go,.NET and PHP
 

Supported(java)

 

Authentication
 

Per-device with SAS token
 

x.509 certificate client
authentication,IAM service

Google IAM
Cloud Resource Manager
Google signing
 

OAuth2 ,
Mutual SSl

 

Communication Patterns
 

Telemetry, Command
 

Telemetry, Command
 

Telemetry, Command
 

Telemetry command

 

Backup
 

Azure backup
 

Use 3rd party tools
,  Aws Datapipe line ,sync to EBS+
snapshot
 

snapshots
 

Analytics data backup /restore
tool

2

Real time analytics
 

Yes
(Event hub along with azure
stream analytics)
 

yes(Kinesis stream)

Yes
 

Yes

3

Edge computing
 

No
(Azure and third party
services  are providing)

Yes
(AWS Greengrass)
 

No
 

Yes

4

Security
 

TLS(server authentication only)

TLS(mutual authentication)
 

End-to-end security using
assymetric key authentication over TLS

Mutual SSL

5

Device level authentication
 

Yes
 

Yes
 

No

Yes