Introduction to Industry 4.

0 and Industrial Internet of Things

Prof. Sudip Misra
Department of Computer Science and Engineering
Indian Institute of Technology, Kharagpur

Lecture - 24
Business Models and Reference Architecture for IIoT: Reference Architecture –
Part 1

In this module, we have talked about the business models that could be used for
Industrial Internet and Industrial IoT scenarios. In this particular lecture, we are going to
talk about the difference architecture. In the context of the business, we have talked
about the possible business models, that could be used in the previous lectures in this
particular module, but we should be also able to use the different architectural platforms
for transformation of the business for the adoption of Industrial IoT. Let us look at the
IIoT reference architecture.

(Refer Slide Time: 01:07)

IIoT reference architecture is governed by the Industrial Internet Reference Architecture.

This is sort of a standard architecture that IIoT systems globally tend to use and tend to
follow. IIRA - Industrial Internet Reference Architecture is the architectural standard,
that is used for most of these IIoT applications in these industries. So, it is a standard
based architecture, which was proposed by the Industrial Internet Consortium – IIC, their
technology working group came up with this architecture. The current version of it is
IIRA v1.8. So, 1.8 is the current version of the Industrial Internet Reference
Architecture. And this architecture broadly applies in the industrial systems to allow
interoperability, mapping application technologies, and in guiding the use of these
technologies by different application users.

(Refer Slide Time: 02:14)

So, for the IIRA one of the central thematic concerns is the safety. And many of these
architectural elements basically have been designed in such a way that safety has been
kept in mind. IIRA safety concern is important and after safety is the issue of security.
Security in terms of everything, security of information, security of data security of the
systems and so on. Safety and security are paramount and central thematic considerations
while coming up with the IIRA architecture for use with IIoT industrial applications.

In terms of safety, it is important to know for the system that is operating, what is the
condition of it. It is also important to ensure that there is no unexpected risk of physical
damage or injury to people due to some malfunctioning or potentially malfunctioning
systems in the future. And third safety concern is to ensure that there is no damage to
property or environment, due to any kind of malfunctioning now or in the future.
(Refer Slide Time: 03:51)

These are some of these KPIs for occupational safety and health. These KPIs are
basically a measure of the activities of an organization. These KPIs also help measure the
connectivity and communication with customers, providing valuable feedback to the
customers, and driving towards further improvement; there are different KPIs.

(Refer Slide Time: 04:20)

So, these KPIs based on their types can be categorized into two types. One is the leading
KPI, and it is mainly used to predict the economy, it is input-oriented, and is hard to
measure. And as the name suggests the latter category is the lagging KPI, which is a
technical indicator, which changes after the economy has initiated. So, unlike the leading
KPI, which is input-oriented, the lagging KPI is output-oriented, and this is hard to
improve. Leading KPI talks about predicting, predicting the economy. And lagging KPI
talking about after the economy has started to improve. So, basically these two
perspectives are different. Leading KPI and lagging KPI are the two different KPI
categories for OSH.

(Refer Slide Time: 05:25)

So, leading KPIs for example, percentage of managers with adequate operational and
safety, health safety training, percentage of workers with adequate operational and safety
training, then, frequency of observed unsafe behavior, number of OSH audits, these are
some of these different KPIs under the leading KPIs category. Under the lagging KPIs
category, we have number of lost-time incident frequency rate, how many times the
different incidents have occurred, and we have lost time due to those incidents, and the
frequency of occurrence of those incidents. Production days lost due to sickness absence,
incidents or near misses, and number of fatalities, these are the lagging KPIs.
(Refer Slide Time: 06:22)

So, I was talking about the IIC, which is the Industrial Internet Consortium, which is a
non-profit consortium doing different things promoting open standards, standards for
interoperability, supporting architectures for promoting interpretability across different
technologies. And these technologies open standards, interoperable standards these are
used in industries and for M2M communication, in these industries.

IIC is the one which came up with that the Industrial Internet Reference Architecture,
IIRA architecture. Industrial Internet Consortium - IIC has also among its focus areas
and activities has the development of Testbeds for different trials also a major concern.
So, testbeds are an area of major focus and activity of the different IIC members.
(Refer Slide Time: 07:33)

So, Industrial Internet Consortium continues to do different activities, tasks and so on,
they have their different working groups. So, these working groups do different things
they innovate, they come up with different opportunities of the use of new technologies,
new applications, new processes, new products, new services, etcetera, which could be
initiated, conceptualized, and could be rigorously tested, in the different testbeds that I
just talked about a while back.

And these could be done before they are actually launched in the market. These are the
ones these are the different activities that are promoted by the IIC.
(Refer Slide Time: 08:16)

So, II Reference Architecture (IIRA) - Industrial Internet Reference Architecture, IIRA

has different architectural components, which are part of their framework, number one is
the stakeholder. So, we need to understand some of these different concepts. So, number
one is the stakeholder stakeholders are individuals teams or organizations having interest
concerning the system.

The next concept is linked to the stakeholders; these are the viewpoints. So, these
stakeholders have some interest in the viewpoints and the system. So, viewpoints are
from the stakeholders, which are basically the collection of different ideas from the
stakeholder’s collection, grouping of them, describing these different ideas, analyzing the
ideas, and solving the set of specific concerns.
(Refer Slide Time: 09:13)

This is the IIRA framework. So, as you can see over here let me just explain it to you.
So, this shows that there are different stakeholder groups--stakeholder group 1,
stakeholder group 2. These stakeholders have different framework sorry have different
viewpoints, viewpoint 1.1, viewpoint 1.2, etcetera.

Similarly, stakeholder 2 would have their own viewpoint, which could be again
classified as 2.1, 2.2 etcetera, as the case may be. So, that is the concept of the
viewpoints, which is basically a, a collection of different ideas from different
stakeholders. So, these step these stakeholders come up with different viewpoints and
these viewpoints essentially help in coming up with these different architectural
components architecture 1.1, 1.2 etcetera, and architecture 2, 2.1, 2.2 etcetera, mapping
directly to these different viewpoints.
(Refer Slide Time: 10:24)

Next comes the concept of the frame, the architectural frame, which is a collection of
ways which identify, describe, and analyze the ideas of the different stakeholders. The
concept of architectural representation is important; it is the collection of outcomes of
architectural frame and are expressed as views.

(Refer Slide Time: 10:47)

Different IIoT architecture implementation patterns are have been proposed. So, we will
go through three different patterns, number one is the three-tier architecture pattern,
number two is the gateway mediated edge connectivity, and management architecture
pattern, and third is the layered data bus pattern.

(Refer Slide Time: 11:10)

So, let us go through each of these, to start with we will first go through the three-tier
architecture pattern. In this three-tier architecture pattern, we have three different layers
we have the edge layer, the platform layer and the enterprise layer. So, edge layer
basically talks about all these different edge devices, basically the device is connecting to
these different gateways at the edge and so on.

And these devices and the gateways etcetera are going to send the data to the platform
layer, these are this is the platform layer, where the data are going to be analyzed, and so
on. And further the data are going to be processed and sent to the enterprise layer. The
control flow is in the reverse direction, data could flow from, the edge layer to the
platform layer directly, or could directly also flow to the enterprise layer.
(Refer Slide Time: 12:17)

So, this three-tier architecture pattern let us go through the different components, in
further more detail. So, we have the edge layer, which gathers data from the edge nodes,
the architecture includes the breadth of distribution, governance, and location of the data.
Breadth of distribution of the data, governance of the data, and the location of the data
meaning that where from the data are coming from these individual devices. Platform
layer basically it is concerning receiving, processing, and forwarding control commands
from the enterprise layer to the edge layer.

(Refer Slide Time: 12:56)

And finally, the enterprise layer concerns receiving data flows from the edge layer and
the platform layer. So, I mean directly from the edge layer, the enterprise layer could be
receiving the data or it could be from the platform layer, which again receives the data
from the edge layer. So, the enterprise layer basically implements domain-specific
requirements, decision support systems, and other requirements such as interfaces to end-
users.

(Refer Slide Time: 13:25)

The next type of architecture is the gateway-mediated edge architecture. Here basically
we are talking about incorporation of the, the gateway, gateway concept. So, we have all
these different sensors and actuators, these sensors and actuators throw in lot of data.
And through the different gateway devices, edge devices, gateway devices and so on the
data are sent to the wide area network for further dissemination of different information
and so on.
(Refer Slide Time: 14:04)

So, like the previous architecture here again let us go through each of these architectural
components of the gate gateway-mediated edge architecture, in further detail. So, this
gateway mediated edge architecture consists of a local area network for the IIoT edge
system and the gateway connecting to the wide area network. The local area network
may use a hub and spoke topology or, a mesh topology.

Mesh topology is one where the different nodes are connected to one another in the form
of a mesh. So, there are lot of different links redundant and many different links are
there. So, basically what happens is if one link has broken, still, then the different nodes
may not be completely disconnected from one another, because there might be some
alternate links directly or indirectly, which will basically handle the problem.

So, the other one is the hub-and-spoke topology, which is very similar to the star
topology where you have the hub and there are different spoke are connected to the
different other edge nodes and together, you have a central kind of hub device
connecting to these different other IoT devices, other edge devices. So, this becomes a
star kind of topology or a hub-and-spoke topology.
(Refer Slide Time: 15:30)

So, you have this local area network, which supports two different types of topologies,
the mesh topology and the hub-and-spoke network topology. Then you, we have these
gateway devices acting as the management points for the edge devices locally
transferring helping in the transferring of the data, processing of the data, and analysis of
the data.

(Refer Slide Time: 15:45)

The third type of architectural pattern is the layered databus pattern. So, one thing I
should mention before we go any further is all these different architectural patterns are
sort of like common templates, which could be used in order to, which could be used as
reference architectures in order to deploy IIoT solutions in the industries.

So, now let us go through this third architectural pattern, which is the layered databus
pattern. So, here we have three different tiers, we have the smart machines, then we have
system of systems, and the industrial internet. So, a system of systems is basically a
complex system consisting of different, different subsystems together working together
generating lot of complex, executing lot of complex algorithm, generating lot of data
handling to be handled, and processed to get lot of insights about what is going on down
underneath. This is what is done by the layered databus.

And then you have the industrial internet. We have these different applications running
on the industrial internet in different industrial application sites, same site, same site in
the same industry different groups or different, different sites of the same company or it
could be that different companies are interconnected together. So, this is the layered data
bus layered data bus pattern, which is different from the previous two patterns that we
have just discussed.

(Refer Slide Time: 17:30)

Let us go through this pattern also little bit, in further detail. So, here we are talking
about use of smart machines. So, these smart machines at the lowest layer will help in
local control and automation processes. System of systems layer allows complex systems
to be executed, complex monitoring, complex processing, and analytic applications, all
of these things could be executed at this particular layer.

(Refer Slide Time: 17:52)

Then you have the layered data bus pattern, which is applicable in the field of control
local monitoring and analytics, and the database basically communicates between the
applications and devices through this communication is enabled with the help of the data
bus. This communication will essentially help in allowing interoperable communication
between different endpoints, between different machines.

(Refer Slide Time: 18:18)

Layered databus pattern basically allows first device to device integration with minimum
response time, automatic data and application delivery, scalable integration of devices,
availability of the system making, it higher and hierarchical subsystem isolation. So, this
isolation part is very important subsystem isolation, because this will help in making the
solutions much more scalable and adaptable, adaptable to what; adaptable to different
changes.

(Refer Slide Time: 19:00)

With this we come to the end of the first part of the reference architecture. So, we have
seen that this IIRA reference architecture is sort of like a de facto kind of standard being
used for the implementation of IIoT in the industries. And this connects well with the
business model of this particular unit in the business model, we have seen the different
types of business models that could be adopted and this reference architecture is the one
which takes it further technically. This reference architecture is sitting between the
business issues and the technicalities. And it is trying to map the business requirements
with the technical requirements by providing different, different architectural solutions,
common architectural solutions, which could be adopted in order to deploy IIoT
solutions in the industries. These are some of these references, which you could use and
go through further. You could also dig on your own to find out more different other
differences on this topic.

Thank you.