Department of Software Engineering,

Traditional Data Flow in IoT

IoT Network Core Network

Image Source: http://events17.linuxfoundation.org/sites/events/files/slides/Intelligence%20at%20the%20Edge.pdf

2
 What is Architectural Plan?
Smart
 In present days, networks Home
run the modern business

 So, it should never be built

without careful planning

 Architecture is how you design (i.e. graphical Driving forces:

structure) your application or solution.  Scale
 Security
 Essence of IoT architecture:  Constrained devices
 how the data is  Massive data
 transported,
 collected,  Data analysis
 analyzed, and  Support to legacy
 ultimately acted upon. devices

3
 Basic 3-Layer architecture
• Perception layer is the physical layer,
which has sensors for sensing and
gathering information about the
environment.

• Network layer is responsible for

connecting to other smart things,
network devices, and servers. Its
features are also used for transmitting
and processing sensor data.

• Application layer is responsible for

delivering application specific services
to the user.
– For example, smart homes, smart
cities, smart health, etc.

Source: https://www.pinterest.com/pin/641129696942354756/

4
Emergence of Standard IoT Architecture
• However, the basic 3-layer architecture did not address many issues.
– e.g. Compatibility, Safety & Security, Reusability, Heterogeneity, etc.

• In fact, the IoT did not have any standard defined architecture of working which is strictly
followed universally.

Few Issues:

 Highly fragmented marked with limited vendor

specific applications

 Each silo contains its own technologies

without interoperability

 Incompatibility for seamless integration

between heterogeneous applications and
devices

• So, in the past several years, architectural standards and frameworks have emerged
• Two best-known architectures:
– oneM2M architecture
– IoT World Forum architecture
Source: https://onem2m.org/using-
5
onem2m/developers/basics
 oneM2M Architecture
Goal of oneM2M architecture:
• to create a common services layer, which can be readily embedded in field
devices to allow communication with application servers.

Challenges in IoT Architecture: • Using the smart building use case, a

• heterogeneity of devices, security application can detect when
• heterogeneity of software, nobody is in the building.
• Heterogeneity of access methods • It could then trigger lights to be switched
off and for the HVAC system to operate
on a reduced setting.
Image Source: https://onem2m.org/using-onem2m/developers/basics

6
oneM2M Architecture
• Proposed by Européen Télécommunications Standards Institute (ETSI)
• oneM2M architecture divides IoT functions into three major domains.
First

• Defines application-layer protocols

• Attempts to standardize northbound API

• API stands for Application Programming Interface.

Note: Interface can be thought of as a contract of

service between two applications or parties. This
contract defines how the two communicate with each
other using requests and responses.

• A northbound interface allows a particular

component of a network to communicate with a
higher-level component.

• Applications have their own sets of data models

7
oneM2M Architecture
• Proposed by Européen Télécommunications Standards Institute (ETSI)
• oneM2M architecture divides IoT functions into three major domains.
First Second
• horizontal framework
across the vertical
industry applications.

• Include:
• the physical network
that the IoT
applications run on.
(e.g. backhaul
network)
• the underlying
management
protocols
• the hardware
cont…
8
oneM2M Architecture
• Proposed by European Telecommunications Standards Institute (ETSI)
• oneM2M architecture divides IoT functions into three major domains.
First Second
• top is the common
services layer

• This layer adds APIs and

middleware supporting
third-party services and
applications.

• Service layer can be

readily embedded within
various hardware and
software nodes

• A RESTful API uses HTTP

requests to GET, PUT,
POST and DELETE data.
9
oneM2M Architecture
• Proposed by European Telecommunications Standards Institute (ETSI)
• oneM2M architecture divides IoT functions into three major domains.
First Second Third

IoT devices, &

communication
network
(802.15.4, LoRa,
WiFi)

10
IoTWF Architecture – 7 Layer Stack
• IoTWF architectural committee (led by Cisco, • Control flowing from the
IBM, Rockwell Automation, and others) center to the edge

• Decompose the IoT

problem into smaller parts

• Identify different
technologies at each layer

• Different parts of a system

can be provided by
different vendors

• Tiered security model

enforced at the transition
points between levels
• offers a clean, simplified perspective on IoT
• includes edge computing, data storage, and access • Define interfaces that
• succinct way of visualizing IoT from a technical perspective leads to interoperability

11
 Layers 1 & 2
Layer 1: Physical Devices Layer 2: Connectivity Layer
and Controllers Layer • focus is on connectivity

• home of the “things” in IoT

• “things” can be from a

microscopic sensors to giant
machines in a factory

• primary function is
generating data

• capable of being queried

and/or controlled over a
network.

12
 Layer 3 : Fog Layer
Basic principle:

Layer 3: Fog/Edge Computing Layer information processing is initiated as

early and as close to the edge of the
• often referred to as the “fog” layer network as possible.
• emphasis is on
– Data reduction by filtering and
cleaning up

– Reformatting and compressing data

– Initial processing of data (e.g. alert

generation, data validation, etc)

13
 Layer 3 : Fog Layer
-Another important function that
occurs at Layer 3 is the evaluation of
data to see if it can be filtered or
aggregated before being sent to a
higher layer.
This also allows for data to be
reformatted or decoded, making
additional processing by other
systems easier. Thus, a critical
function is assessing the data to see
if predefined thresholds are crossed
and any action or alerts need to be
sent

14
 Upper Layers: Layers 4–7

Layers Functions
Layer 4: Data Accumulation • Captures data and stores it for applications
• Convert event-based data to query-based processing
Layer 5: Data Abstraction • Reconciles multiple data formats
• Ensures consistent semantics for various data sources
• Confirmation about dataset completeness
Layer 6: Application • Interpret data using software applications
• Applications may monitor, control, and provide report
based on analysing the data
Layer 7: Collaboration and • Consumes and shares the application information
processes • Collaborating and communicating IoT information
15
 Simplified IoT Architecture
• It highlights the fundamental building blocks that are common to most IoT
systems and which is intended to help in designing an IoT network.

• IoT architectural framework is presented as two parallel stacks

• Core IoT Functional Stack
• IoT Data Management and Compute Stack

16
IoT Data Management and Compute Stack

(1) Cloud computing layer

 prediction by Cisco that by 2020 there will be more than 50 billion

devices connected to some form of an IP network. Clearly, the data
generated by these devices must also be of serious concern.
 In sensor networks, the vast majority of data generated is
unstructured
 For example, the majority of data generated by a smart meter is
nothing more than polling data; the communications system simply
determines whether a network connection to the meter is still active.
This data on its own is of very little value. The real value of a smart
meter is the metering data read by the meter management system
(MMS)

17
IoT Data Management and Compute Stack

(1) Cloud computing layer

 These new requirements include the following
Minimizing latency
 Milliseconds matters, to prevent manufacturing line shutdowns or
restore electrical service. Analyzing data close to the device that
collected the data can make a difference between averting disaster
and a cascading system failure.

Conserving network bandwidth

 Offshore oil rigs generate 500 GB of data weekly. Commercial jets
generate 10 TB for every 30 minutes of flight. It is not practical to
transport vast amounts of data from thousands or hundreds of
thousands of edge devices to the cloud. Nor is it necessary because
many critical analyses do not require cloud-scale processing and
storage.
18
IoT Data Management and Compute Stack

(1) Cloud computing layer

 These new requirements include the following
Increasing local efficiency
 Collecting and securing data across a wide geographic area with
different environmental conditions may not be useful. The
environmental conditions in one area will trigger a local response
independent from the conditions of another site hundreds of miles
away. Analyzing both areas in the same cloud system may not be
necessary for immediate efficiency.

19
IoT Data Management and Compute Stack

(1) Cloud computing layer

 In smart meters, the processing location is outside the smart object. A

natural location for this processing activity is the cloud.
 Smart objects need to connect to the cloud, and data processing is
centralized.
 One advantage of this model is simplicity. Objects just need to
connect to a central cloud application.
 That application has visibility over all the IoT nodes and can process all
the analytics needed today and in the future.

20
IoT Data Management and Compute Stack

(1) Cloud computing layer

21
IoT Data Management and Compute Stack

(2) Fog Computing layer

 to distribute data management throughout the IoT system, as close
to the edge of the IP network as possible

 Any device with computing, storage, and network connectivity can

be a fog node. Examples include industrial controllers, switches,
routers, embedded servers, and IoT gateways.

 The fog layer thus provides a distributed edge control loop capability,
where devices can be monitored, controlled, and analyzed in real
time without the need to wait for communication from the central
analytics and application servers in the cloud.

22
IoT Data Management and Compute Stack

(2) Fog Computing layer

 For example, there might be a fog router on an oil derrick that is
monitoring all the sensor activity at that location. Because the fog
node is able to analyze information from all the sensors on that
derrick, it can provide contextual analysis of the messages it is
receiving and may decide to send back only the relevant information
over the backhaul network to the cloud.

 In this way, it is performing distributed analytics such that the

volume of data sent upstream is greatly reduced and is much more
useful to application and analytics servers residing in the cloud.

23
IoT Data Management and Compute Stack

(2) Fog Computing layer

 For example, tire pressure sensors on a large truck in an open-pit
mine might continually report measurements all day long. There may
be only minor pressure changes that are well within tolerance limits,
making continual reporting to the cloud unnecessary.

 Is it really useful to continually send such data back to the cloud over
a potentially expensive backhaul connection? With a fog node on the
truck, it is possible to not only measure the pressure of all tires at
once but also combine this data with information coming from other
sensors in the engine, hydraulics, and so on. With this approach, the
fog node sends alert data upstream only if an actual problem is
beginning to occur on the truck that affects operational efficiency.

24
IoT Data Management and Compute Stack

(3) Edge Computing layer

 it resides directly in the sensors and IoT devices.
 IoT devices and sensors often have constrained resources, however,
as compute capabilities increase. Some new classes of IoT endpoints
have enough compute capabilities to perform at least low-level
analytics and filtering to make basic decisions.
 For example, consider a water sensor on a fire hydrant. While a fog
node sitting on an electrical pole in the distribution network may
have an excellent view of all the fire hydrants in a local
neighborhood, a node on each hydrant would have clear view of a
water pressure drop on its own line and would be able to quickly
generate an alert of a localized problem. The fog node, on the other
hand, would have a wider view and would be able to ascertain
whether the problem was more than just localized but was affecting
the entire area.
25
IoT Data Management and Compute Stack

Edge Computing layer

 Another example is in the use of smart meters. Edge compute–
capable meters are able to communicate with each other to share
information on small subsets of the electrical distribution grid to
monitor localized power quality and consumption, and they can
inform a fog node of events that may pertain to only tiny sections of
the grid. Models such as these help ensure the highest quality of
power delivery to customers.

26
Lessons Learned

 What is the need of IoT architecture?

 Different type of IoT Architectures

 About oneM2M architecture

 About IoT WF architecture

 About Simplified IoT Architecture

27
 Figures and slide materials are taken from the following sources:

1. David Hanes et al., “IoT Fundamentals: Networking Technologies, Protocols,

and Use Cases for the Internet of Things”, 1st Edition, 2018, Pearson India.

03-08-2023 Dr. Manas Khatua 28