EL

IIoT Analytics and Data Management: Introduction

PT Dr. Sudip Misra

Professor
Department of Computer Science and Engineering
Indian Institute of Technology Kharagpur
Email: [email protected]
Website: http://cse.iitkgp.ac.in/~smisra/
Research Lab: cse.iitkgp.ac.in/~smisra/swan/

Industry 4.0 and Industrial Internet of Things 1

IIoT Analytics : Necessity

EL
PT
Industry 4.0 and Industrial Internet of Things 2
IIoT Analytics : Definition

EL
PT
Industry 4.0 and Industrial Internet of Things 3
IIoT Analytics : Definition (cont.)

EL
PT
Dynamic operations optimization
Prognostic maintenance
Real-time data analysis

Industry 4.0 and Industrial Internet of Things 4

IIoT Analytics : Types

EL
IIoT Analytics

PT
Descriptive
Analytics
Predictive
Analytics
Prescriptive
Analytics

Industry 4.0 and Industrial Internet of Things 5

IIoT Analytics : Challenges

EL
PT
Industry 4.0 and Industrial Internet of Things 6
IIoT Analytics: Data Science

EL
 Big Data Analytics
 Volume, velocity, variability, veracity, variety
 Industrial automation, system health monitoring, predictive maintenance,
remote monitoring

PT
 Artificial Intelligence
 Deep Learning (DL)
 Machine Learning (ML)
Instead of physics-based models, ML and DL enable a data-driven system
modelling approach.

Industry 4.0 and Industrial Internet of Things 7

IIoT Analytics: Machine Learning

EL
 May be supervised or unsupervised

PT
 Feature extraction: Convert raw data to information that relates to the
physical state of the asset.
 Supervised algorithms are useful when it is feasible to acquire training
data for the different states (or classes) that need to be modeled.
 Unsupervised methods do not use labeled data. These algorithms find
structure in the input data on its own.

Industry 4.0 and Industrial Internet of Things 8

IIoT Analytics: Deep Learning

EL
 Feature engineering is absent





PT
Raw data from sensors is directly fed into deep learning algorithms
Learns features automatically without manually specifying them
Largely based on neural networks
Require large amount of computation power (GPU servers)
Gaining popularity due to advances in computing

Industry 4.0 and Industrial Internet of Things 9

IIoT Analytics: Timescales

EL
 Baseline analytics (ms)
 Detect irregular behavior of
asset quickly
 Diagnostic analytics (min)

PT
 Identify root cause of anomaly
 Prognostic analytics (hrs)
 Inform about remaining useful
life of an asset

Industry 4.0 and Industrial Internet of Things 10

IIoT Analytics: Deployment

EL
 Deployment of analytics typically consists of three steps:
 train a (predictive) analytics model
 test and validate the model on previously unseen data
 deploy the model to make predictions on real (streaming) data.

PT
Industry 4.0 and Industrial Internet of Things 11
IIoT Analytics: Real-time

EL
 Streaming real-time analytics
 Most often used for IoT processing
 Take action immediately on some event with the source
 Send out alert on temperature sensor reaching a threshold
 Send out notification about low tire pressure in smart car

PT
 Generating instant alerts requires stream processing. Process events in
real-time to match a predefined set of rules
 Edge processing, data aggregation and down-sampling
 Complex Event Processing software such as Apache Storm, Esper etc may
be used.

Industry 4.0 and Industrial Internet of Things 12

IIoT Analytics: Batch Processing

EL
 Batch-oriented analytics
 Improve accuracy of the streaming layer analytics
 Useful for long-term statistics
 Average temperature in room for last month

PT
 Power usage of house in last year
 Distributed analytics: Batch processing can be used to get a better
overall picture by aggregating data sources from geo-distributed
sources.
 Software such as Apache Hadoop and Apache Spark may be used

Industry 4.0 and Industrial Internet of Things 13

IIoT Analytics: Model Building

EL
 Anomaly detection
(e.g., Gaussian Mixture
Model)
 Classification

Machine)

PT
(e.g., Support Vector

 Regression
(e.g., Bayes Regression)

Industry 4.0 and Industrial Internet of Things 14

IIoT Analytics: Value Drivers

EL
 New revenue streams
 Upgrading existing products
 Changing the business model
 Creating new business models

PT
 Reduce costs
 Data-driven process optimization
 Data-driven process automation
 Data-driven product optimization

Industry 4.0 and Industrial Internet of Things 15

IIoT Analytics: Applications across the value chain

EL
 R&D: Analyze product usage characteristics and feed back generated data
to improve the product in the next cycle
 Manufacturing / Operations
 Predictive maintenance
 Decision support systems for industrial processes

PT
 Optimizing machine parameters: Correlated cause and effect parameters such as
machine speed
 Logistics / Supply chain
 Supply chain optimization: forecast shortages, reduce overall inventory levels etc
 Fleet management: optimize to reduce transportation and fuel cost
 Marketing and Sales: Propose suitable upgrades as per user behavior

Industry 4.0 and Industrial Internet of Things 16

IIoT Analytics: Aircraft Example

EL
PT
Industry 4.0 and Industrial Internet of Things 17
 References

EL
[1] N. Anderson, W.W. Diab, T. French, K.E. Harper, S. Lin, D. Nair, W. Sobel, “The Industrial Internet of Things
Volume T3: Analytics Framework”, White Paper, Industrial Internet Consortium, 2017.
[2] IoT Analytics Inc., “Industrial Analytics 2016/2017: The current state of data analytics usage in industrial
companies”, Tech Report, 2016.
[3] A. Minteer, “Analytics for the Internet of Things (IoT)”, Packt Publishing, 2017
[4] S. Verma, Y. Kawamoto, Z.M. Fadlullah, H. Nishiyama, N. Kato, “A Survey on Network Methodologies for

PT
Real-Time Analytics of Massive IoT Data and Open Research Issues “, IEEE Communication Surveys & Tutorials,
vol. 19, no. 3, pp. 1457–1477, 2017.

Industry 4.0 and Industrial Internet of Things 18

 EL
PT
Introduction to Internet of Things 19
 EL
IIoT Analytics and Data Management:
Machine Learning and Data Science – Part 1

PT Dr. Sudip Misra

Professor
Department of Computer Science and Engineering
Indian Institute of Technology Kharagpur
Email: [email protected]
Website: http://cse.iitkgp.ac.in/~smisra/
Research Lab: cse.iitkgp.ac.in/~smisra/swan/

Industry 4.0 and Industrial Internet of Things 1

 EL
Basics of Machine Learning

PT
Industry 4.0 and Industrial Internet of Things 2
 What is Machine Learning?

EL
Machine learning is a subset of Artificial Intelligence which enables
machines to make decisions based on their experience rather than being

PT
explicitly programmed.

Source: Google Cloud AI Adventures, figure redrawn from URL:https://towardsdatascience.com/what-is-machine-learning-8c6871016736

Industry 4.0 and Industrial Internet of Things 3

 What is Data Science?

EL
Computer Science

Statistical Data
Research Processing

PT
Data
Science
Mathematics Domain
Machine Expertise
Learning

Source: Quora
URL: https://www.quora.com/What-is-data-science

Industry 4.0 and Industrial Internet of Things 4

 Machine Learning

EL
Using data to answer questions

Training Prediction
Using Data Answer questions

PT
Source: Google Cloud AI Adventures, figure redrawn from URL:https://towardsdatascience.com/what-is-machine-learning-8c6871016736

Industry 4.0 and Industrial Internet of Things 5

 How Machine Learning works?

EL
Labeled or Training
Unlabeled
Dataset
Created
Machine Learning
Model
Algorithm

PT New Input
Data
Testing

Serve
Predictions
Predicted
Output

Source: Google Cloud AI Adventures, figure redrawn from URL:https://towardsdatascience.com/what-is-machine-learning-8c6871016736

Industry 4.0 and Industrial Internet of Things 6

Types of Machine Learning Algorithms

EL
Machine Learning

1. Unsupervised Learning

PT 2. Supervised Learning
3. Reinforcement Learning

Industry 4.0 and Industrial Internet of Things 7

 1. Unsupervised Learning

EL
This machine learning technique is used to identify similar groups of data,
coined as clustering. The segregation of data is performed on unlabeled
dataset, based on the inner structure of the data without looking into the
specific outcome.

Clustering

PT
Data points belong to
completely one cluster
Ex. K-Means
Hard Clustering Soft Clustering
Data points may
belong to more than
one cluster
Ex. Fuzzy c-Means
Source: analyticsvidya website URL: https://www.analyticsvidhya.com/blog/2016/11/an-introduction-to-clustering-and-different-methods-of-clustering/

Industry 4.0 and Industrial Internet of Things 8

Difference between K-Means and Fuzzy c-Means Algorithm

EL
K- Fuzzy c-
Means Means

1. One data point may 1. One data point may belong to

belong to only one cluster more than one cluster

PT
2. K-Means may not be as
fast as FCM
2. FCM is extremely faster than
K-Means

Industry 4.0 and Industrial Internet of Things 9

2. Supervised Learning Algorithm

EL
It is used to classify the dataset by learning the mapping function from the
labeled dataset.

Supervised
Learning

PT
When the
output variable
is a real number
such as dollars
or weight
Regression Classification
When the output
variable is a
category such as
red or blue

Industry 4.0 and Industrial Internet of Things 10

Regression : Linear Regression

EL
It is supervised learning problem which learns a linear function from the
given instances of X (independent variable) and Y (dependent variable)
values, so that it can predict Y for an unknown X.

Y Y=B0X+B1X+e

PT
B1
Y intercept Population Slope

B0

Industry 4.0 and Industrial Internet of Things 11

 Classification: Decision Tree

EL
 Tree-based machine learning algorithm Outlook
used for classification sunny rain
overcast

 Non-linear function with two types of Humidity Wind

nodes: decision nodes and leaf nodes yes
weak
high normal strong

PT
 Decision node is used to test or decide
yes no
the outcome based on some value of yes no

an attribute

 Leaf node denotes the classification of

an example
Figure redrawn from URL: https://nullpointerexception1.wordpress.com/2017/12/16/a-tutorial-to-understand-decision-tree-id3-learning-algorithm/

Industry 4.0 and Industrial Internet of Things 12

 3. Reinforcement Learning Algorithm

EL
It is a machine learning algorithm which enables machines to improve
its performance by automatically learning the ideal behaviors for a
specific environment.

State
Environment

PT
Reward
State
Enviornment
Action Agent
Reward
Action
Source: “Learn Unity ML-Agents – Fundamentals of Unity Machine Learning” by Micheal Lanham

Industry 4.0 and Industrial Internet of Things 13

RL – An Analogy

EL
Environment Environment
Observe Observe

PT
Fire Water

Industry 4.0 and Industrial Internet of Things 14

 Differences between RL and Supervised Learning

EL
Reinforce Super-
ment vised
Learning Learning
1. Here agent is guided by an
1. There is no external supervisor
external supervisor who has the
to guide the agent.
knowledge of the environment.

PT
2. No problem faced during the 2. Problem faced during the
circumstances. The agent has circumstances. The agent has many
many combinations of subtasks to combinations of subtasks to
achieve the objective. achieve the objective.
3. There is a reward function
which acts as a feedback to the 3. There is no reward function.
agent.

Source: Analytics Vidhya URL:https://www.analyticsvidhya.com/blog/2017/01/introduction-to-reinforcement-learning-implementation/

Industry 4.0 and Industrial Internet of Things 15

 Difference between RL and Unsupervised Learning

EL
Reinforce Unsuperv
ment ised
Learning Learning

1. There is a mapping between 1. There is no mapping

input and output. between input and output.

PT
2. It builds a knowledge graph 2. It finds the underlying
from the constant feedbacks of pattern.
the corresponding actions.

Source: Analytics Vidhya URL:https://www.analyticsvidhya.com/blog/2017/01/introduction-to-reinforcement-learning-implementation/

Industry 4.0 and Industrial Internet of Things 16

 EL
Integration of Machine Learning with IIoT

PT
Industry 4.0 and Industrial Internet of Things 17
Various industries utilizing IIoT with Machine Learning

EL
Retail Travel

Social

PT
Health
Finance media
care

Industry 4.0 and Industrial Internet of Things 18

 Applications of IIoT with Machine Learning (Contd…)

EL
 Pfizer exploits IBM Watson for drug
discovery
Healthcare

for patients
PT
 Genentech provide personalized treatment

Source: Top 10 Industrial Applications of Machine Learning

URL: https://www.dezyre.com/article/top-10-industrial-applications-of-machine-learning/364

Industry 4.0 and Industrial Internet of Things 19

 Applications of IIoT with Machine Learning (Contd…)

EL
 Fraud detection
Finance

PT
 Targeting focused account holders

Source: Top 10 Industrial Applications of Machine Learning

URL: https://www.dezyre.com/article/top-10-industrial-applications-of-machine-learning/364

Industry 4.0 and Industrial Internet of Things 20

 Applications of IIoT with Machine Learning (Contd…)

EL
 Product recommendation
Retail
 Improved customer service

PT
Source: Top 10 Industrial Applications of Machine Learning
URL: https://www.dezyre.com/article/top-10-industrial-applications-of-machine-learning/364

Industry 4.0 and Industrial Internet of Things 21

 Applications of IIoT with Machine Learning (Contd…)

EL
 Dynamic price setup
Travel
 Sentiment analysis to act as trip advisor

PT
Source: Top 10 Industrial Applications of Machine Learning
URL: https://www.dezyre.com/article/top-10-industrial-applications-of-machine-learning/364

Industry 4.0 and Industrial Internet of Things 22

 Applications of IIoT with Machine Learning (Contd…)

EL
 Facebook uses ANN for tagging
faces Social
media

PT
 LinkedIn uses machine learning
technology for suggesting job

Source: Top 10 Industrial Applications of Machine Learning

URL: https://www.dezyre.com/article/top-10-industrial-applications-of-machine-learning/364

Industry 4.0 and Industrial Internet of Things 23

 Instances of IIoT with Machine learning

EL
 Perform complex analytical process

 Deliver real-time perception

ThingWorx platform

PT
 Ability of condition monitoring

 Ability of predictive analytics and recommendation

Source: Deliver Industrial IoT Analytics with ThingWorx

URL: https://www.ptc.com/en/products/iot/thingworx-platform/analyze

Industry 4.0 and Industrial Internet of Things 24

 Instances of IIoT with Machine learning (Contd…)

EL
 Help oil and gas engineers to access real
time data and predict anomalies
Toumetis
 Making more advanced smart home

PT
automation

Source: Toumetis URL: https://toumetis.com

Industry 4.0 and Industrial Internet of Things 25

 References-I

EL
[1] Google cloud AI Adventures
URL:https://towardsdatascience.com/what-is-machine-learning-8c6871016736
[2] An introduction to clustering and different methods of clustering
URL: https://www.analyticsvidhya.com/blog/2016/11/an-introduction-to-clustering-and-different-methods-of-
clustering/
[3] Analytics Vidhya
URL:https://www.analyticsvidhya.com/blog/2017/01/introduction-to-reinforcement-learning-implementation/
[4] M. Lanham (2018) Learn Unity ML-Agents – Fundamentals of Unity Machine Learning. Packt publishing

PT
[5] Deep Reinforcement Learning Demystified
https://medium.com/@m.alzantot/deep-reinforcement-learning-demystified-episode-0-2198c05a6124
[6] Top 10 Industrial Applications of Machine Learning
URL: https://www.dezyre.com/article/top-10-industrial-applications-of-machine-learning/364
[7]Toumetis URL: https://toumetis.com
[8] Deliver Industrial IoT Analytics with ThingWorx
URL: https://www.ptc.com/en/products/iot/thingworx-platform/analyze

Industry 4.0 and Industrial Internet of Things 26

 EL
PT
Industry 4.0 and Industrial Internet of Things 27
 EL
IIoT Analytics and Data Management:
Machine Learning and Data Science – Part 2

PT Dr. Sudip Misra

Professor
Department of Computer Science and Engineering
Indian Institute of Technology Kharagpur
Email: [email protected]
Website: http://cse.iitkgp.ac.in/~smisra/
Research Lab: cse.iitkgp.ac.in/~smisra/swan/

Industry 4.0 and Industrial Internet of Things 1

 AI vs Machine learning vs Deep learning
Artificial Intelligence

EL
Making of intelligent machines
which take decisions on its
Artificial Intelligence
own without being explicitly
programmed

Machine Learning Machine Learning

PT
Subset of AI which can learn
automatically from the object
features
Deep
Learning Deep Learning
Subset of machine learning
which can learn automatically
Source Edureka website by finding the features of the
URL: https://www.edureka.co/blog/what-is-deep-learning object on its own

Industry 4.0 and Industrial Internet of Things 3

 Timeline

EL
Artificial Intelligence
Machine Learning

Deep Learning

Source Edureka website

PT
1950 1960 1970

URL: https://www.edureka.co/blog/what-is-deep-learning
1980 1990 2000 2006 2012 2018

Industry 4.0 and Industrial Internet of Things 3

Limitations of Machine learning

EL
1 ML algorithms are not useful for high dimensional data

2 Features have to be explicitly mentioned

PT Solution
Deep Learning

Industry 4.0 and Industrial Internet of Things 4

 Limitations of Machine learning (cont..)

EL
Deep learning

Performance Machine learning

Source Edureka website

PT Amount of data
Figure redrawn from URL: https://www.edureka.co/blog/what-is-deep-learning
High dimensional data
Feature identification/extraction

Industry 4.0 and Industrial Internet of Things 5

Deep Learning

EL
 It is a subfield of machine learning, capable to learn the right
features by its own, basically mimics the working function of
billions of neurons in our brain.

PT
 Deep learning learns features by own
 Deep learning gives better performance like accuracy, when the amount
of data is huge

Industry 4.0 and Industrial Internet of Things 6

 How it works

EL
Deep learning : deep neural network

 Signals travel between neurons in artificial neural network

 In neural network, each neuron is assigned with weightage value

PT
 A high weighted neuron exerts more effect on next layer than others
 Final layer combines all weighted inputs to emerge with a result

Source: Mathworks website

URL: https://in.mathworks.com/discovery/deep-learning.html

Industry 4.0 and Industrial Internet of Things 7

 How it works (cont..)

EL
Hidden layer
Input layer Output layer

PT
Deep refers to number of hidden layers, deep network can have up to 150 hidden layers
Source: Mathworks website Figure redrawn from URL: https://in.mathworks.com/discovery/deep-learning.html

Industry 4.0 and Industrial Internet of Things 8

Understanding analogies of Deep Learning

EL
Let we want to recognize an apple,
 First check shape if Yes
 Then check color if Yes
 Then check its taste if Yes
Apple

PT So it is a nested hierarchy of concept

Deep learning also follows the concept of nested hierarchy, it breaks the
complex task into simple tasks

Industry 4.0 and Industrial Internet of Things 9

 Difference between machine learning and deep
learning

EL
 Deep learning is an “end-to-end learning”, which extracts
features on its own. On the contrary, in machine learning
features are to be explicitly mentioned.

learning converges PT
 In deep learning performance level often improves as the size
of the data increases, whereas in machine learning, shallow

Source: Mathworks website

URL: https://in.mathworks.com/discovery/deep-learning.html

Industry 4.0 and Industrial Internet of Things 10

 EL
Deep learning in IIoT

PT
Industry 4.0 and Industrial Internet of Things 11
Impacts of deep learning in IIoT

EL
IIoT Deep learning

Improve speed Improve accuracy

PT
 For optimization of manufacturing lines in factories
 For stable operations of energy and transportation system
 For system shutdown in emergency

Industry 4.0 and Industrial Internet of Things 12

 Reason for usefulness of deep learning in IIoT

EL
The most important reasons that have made deep learning so useful recently are:

 Requires large amount of labeled data

PT
 Requires high end computational power

Source: Mathworks website

URL: https://in.mathworks.com/discovery/deep-learning.html

Industry 4.0 and Industrial Internet of Things 13

 Critical requirements of deep leaning in IIoT

EL
The important factors required by deep learning methods in IIoT
for solving critical issues
 Large Quantity of data

Source: TOSHIBA website

PT
 High Quality and accuracy of reliable data

URL: https://www.toshiba-sol.co.jp/en/articles/tsoul/20/001.htm

Industry 4.0 and Industrial Internet of Things 14

 Values provided by deep learning in IIoT

EL
Three values provided by deep learning to customers in various business segments

1 Identification or recognition using cameras, sensors etc

3
Source: TOSHIBA website
PT
Prediction/ Inference of human behavior

Autonomous decision control

URL: https://www.toshiba-sol.co.jp/en/articles/tsoul/20/001.html

Industry 4.0 and Industrial Internet of Things 15

 Deep learning as strength of technology

EL
TOSHIBA is using Collaborative Distributed Deep Learning technology
between edge and cloud

Learning process is performed in cloud for high processing

Source: TOSHIBA website

PT
Inference process is conducted in edge for real time processing

URL: https://www.toshiba-sol.co.jp/en/articles/tsoul/20/001.html

Industry 4.0 and Industrial Internet of Things 16

 Deep learning as strength of technology (Contd…)

EL
 Improve yield and productivity in semiconductor
factory

 Adopted drone navigation control system to find

TOSHIBA damage in power transmission line

Source: TOSHIBA website

PT  Predict behavior of workers in warehouses
through wearable devices

 Forecasting power generation in solar power

system
URL: https://www.toshiba-sol.co.jp/en/articles/tsoul/20/001.html

Industry 4.0 and Industrial Internet of Things 17

 Deep learning as strength of technology (Contd…)

EL
 H2O platform utilizes deep learning platform

 Intel’s innovation Nervana, a deep learning processor

PT
 Zebra medical vision systems is applying deep learning techniques

URL: https://www.h2o.ai/, https://ai.intel.com/, https://www.zebra-med.com/

Industry 4.0 and Industrial Internet of Things 18

References-I

EL
[1] What is Deep Learning? Getting started with Deep Learning
URL: https://www.edureka.co/blog/what-is-deep-learning
[2] What is Deep Learning?
URL: https://in.mathworks.com/discovery/deep-learning.html
[3] Deep learning tutorial for beginners
https://www.kaggle.com/kanncaa1/deep-learning-tutorial-for-beginners

PT
[4] D. L. Poole, A. K. Macworth (2017). Artificial Intelligence. Cambridge University Press
[5] R. Chopra (2012) Artificial Intelligence. S. Chand & Company Pvt. Ltd.
[6] TOSHIBA, URL:https://www.toshiba-sol.co.jp/en/articles/tsoul/20/001.html
[7] H2O, URL:https://www.h2o.ai/, Intel URL: https://ai.intel.com/, Zebra-med https://www.zebra-
med.com

Industry 4.0 and Industrial Internet of Things

19
 EL
PT
Industry 4.0 and Industrial Internet of Things 20
 EL
IIoT Analytics and Data Management:
Cloud Computing in IIoT – Part 1

PT Dr. Sudip Misra

Professor
Department of Computer Science and Engineering
Indian Institute of Technology Kharagpur
Email: [email protected]
Website: http://cse.iitkgp.ac.in/~smisra/
Research Lab: cse.iitkgp.ac.in/~smisra/swan/

Industry 4.0 and Industrial Internet of Things 1

Introduction

EL
 IIoT support for Industry 4.0
 Sensing
 Communication
 Computing


PT
Networking
 Achieves digitization in manufacturing and production process

Source: “Industry 4.0: The Industrial Internet of Things”, Apress, 2016

Industry 4.0 and Industrial Internet of Things 2

IIoT and Big Data

EL
 Digitization Process
 Data acquisition
 Asset management
 Resource management


PT
Knowledge management
 Bulk amount of data due to the time series data streams from
end devices
Source: “Industry 4.0: The Industrial Internet of Things”, Apress, 2016

Industry 4.0 and Industrial Internet of Things 3

Need for Cloud

EL
 Major concern to handle huge amount of data
 Nature of data
 Unorganized


PT
M2M sensor data
 From heterogeneous big number of devices
 Varying data quality

Source: “Industry 4.0: The Industrial Internet of Things”, Apress, 2016

Industry 4.0 and Industrial Internet of Things 4

Need for Cloud (Contd.)

EL
Cloud for IIoT

Scalable and secure

High computational speed for services

PT
data monitoring and analytics

Effective data acquisition

Storage of huge amount
of data

Source: “Industry 4.0: The Industrial Internet of Things”, Apress, 2016

Industry 4.0 and Industrial Internet of Things 5

Cloud Computing – Basics

EL
 Suitable for its scientific and business adaptability
 Fulfills the need of what, when and where solutions
 Secure storage and access

PT
 Supports a coherent, expandable and coordinated business
model
 Supports mobile devices

Source: “NIST Cloud Computing Reference Architecture”, NIST

Industry 4.0 and Industrial Internet of Things 6

Cloud Computing

EL
Monitoring
Applications Analytics
Predictions Decision Making

Services

PT
Platform Database
Text
Run time
Middleware

Block Storage Network

Infrastructure
Compute

Source: “NIST Cloud Computing Reference Architecture”, NIST

Industry 4.0 and Industrial Internet of Things 7

Cloud Computing in IIoT: Services

EL
 Three types of services: SaaS, PaaS and IaaS
 Software-as-a-Service (SaaS)
 Industrial applications with web or program Interface


PT
Subscribe-and-use feature to industry clients with final product
 Everything managed by the service provider
 Ex: Industrial Machinery Catalyst from Siemens is a SaaS for industrial
use

Source: “NIST Cloud Computing Reference Architecture”, NIST

Industry 4.0 and Industrial Internet of Things 8

Cloud Computing in IIoT: Services

EL
 Platform-as-a-Service (PaaS)
 Allows industries for self-development of applications
 Clients have control over the application and the configuration
environment

PT
 EX: Predix (GE), Sentience (Honeywell), and MindSphere (Siemens) are
some industrial PaaS providers
 Software firms like Cumulocity, Bosch IoT, and Carriots offer PaaS for IoT
industries
Source: “NIST Cloud Computing Reference Architecture”, NIST

Industry 4.0 and Industrial Internet of Things 9

Cloud Computing in IIoT: Services

EL
 Infrastructure-as-a-Service (IaaS)
 Access to the servers, network and storage and provisioning
 Clients can use cloud to operate a virtual data center
 Used to deploy PaaS and SaaS


PT
Ex: Microsoft Azure, Google Compute Engine, IBM SmartCloud
Enterprise, Rackspace Open Cloud, Amazon Web Services (AWS), etc.

Source: “NIST Cloud Computing Reference Architecture”, NIST

Industry 4.0 and Industrial Internet of Things 10

Cloud Computing in IIoT: Deployment Models

EL
• Cloud set-up for use of any person or industry
• Virtualized resources are publicly shared
• Examples: Google Compute Engine, Amazon Web Service (AWS),
Microsoft Azure, etc.

• Cloud set-up for a single organization

• Virtualized resources are shared with the client only

PT
• managed by the client itself or a third party
• Highly Secure

• Cloud set-up by two or more unique cloud set-up (private or public)

• Designed to have advantages of both private and public
• Flexibility for data and applications movement between private
and public clouds

Source: “NIST Cloud Computing Reference Architecture”, NIST

Industry 4.0 and Industrial Internet of Things 11

Cloud Computing in IIoT: End-users

EL
 End-users are the industries who actually avail the cloud
services
 Services differ form firm to firm based on their products and
services

PT
 Domain of use for IIoT lies in many areas like Healthcare,
Transportation, Manufacturing plants, Refineries, Mining,
Marine and many more.
Source: “Industry 4.0: The Industrial Internet of Things”, Apress, 2016

Industry 4.0 and Industrial Internet of Things 12

Cloud-Based IIoT Architecture

EL
Industrial Applications

Cloud
Data Acquisition and Analytics

PT
Gateway Devices

Sensors and Devices

zz

Source: Gubbi et al., 2013

Industry 4.0 and Industrial Internet of Things 13

Cloud Computing in IIoT

EL
 Industrial big data storage
 Heavy weight algorithms for data analytics
 Prediction of failures before occurrences

PT
 Device provisioning and configuration remotely
 Real-time device monitoring
 Data privacy and security

Source: “Industry 4.0: The Industrial Internet of Things”, Apress, 2016

Industry 4.0 and Industrial Internet of Things 14

Consumer vs. Industrial IoT Cloud Platforms

EL
 Consumer IoT cloud platform
 Very specific applications for end users
 Modest security
 Cost sensitive

PT
 Industrial IoT cloud Platform
 Large number of data points
 QoS
 Robust security
 Return on investment (ROI) sensitive
Source: “The Future of Industrial IoT”, Industrial Internet Consortium

Industry 4.0 and Industrial Internet of Things 15

Industrial Cloud Platform Providers

EL
Bosch IoT
IBM Watson IoT

Intel IoT Cumulocity

Carriots Uptake
Losant IoT AT&T M2X
Amazon AWS

PT
MS Azure
Cisco Jasper C3IoT
XMPRO
AirVantage Amplia IoT
SAP Hana Cloud
Honeywell Sentience TempoIQ
Bitstew Systems GE Predix

Siemens MindSphere
Meshify ThingWorx

Source: “The List of Industrial Cloud Platform Providers”, Element 14

Industry 4.0 and Industrial Internet of Things 16

Industrial Cloud Platform Providers: Our Discussion

EL
 By industrial companies
 GE Predix
 Siemens MindSphere
 Honeywell

PT
 By Software development firms
 C3 IoT
 Uptake
 Meshify
Source: “Will There Be A Dominant IIoT Cloud Platform?”, Element 14

Industry 4.0 and Industrial Internet of Things 17

Predix

EL
 Platform-as-a-service
 Tracking, management and enhancement of capital
 Defines the organization of the system and subassemblies

PT
components of services
 Enables differentiated functionalities of applications
 Digital Twin Technology for learning, estimating, optimizing
and representation of assets
Source: “Industrial Capabilities of Predix”, Predix

Industry 4.0 and Industrial Internet of Things 18

Predix (Contd.)

EL
IT and OT Data

OT Systems Enterprise Data Asset Data External Data

IIoT Platform

Analytics and Distributed Development

Industrial Data Fabric Intelligence
Machine Learning Environments

PT
Industrial Applications

1
2
3
4
5

Source: “Industrial Capabilities of Predix”, Predix

Industry 4.0 and Industrial Internet of Things 19

MindSphere

EL
 Cloud-based operating system platform for IoT
 Open platform-as-a-service in addition to AWS cloud service
 Brings together IoT data from product, factory, machine and

PT
system to exploit the its prosperity
 Enterprise oriented solutions

Source: “MindSphere The cloud-based, open IoT operating system”, MindSphere

Industry 4.0 and Industrial Internet of Things 20

MindSphere (Contd.)

EL
Features Advantages

Security supports Large system network

Procurement and distribution Ecological industry solutions

PT
Extensive analysis and innovations
Various APIs for analysis
along with Digital twins

Automated exploitation of
PaaS with cloud services
performance and intelligence

Source: “MindSphere The cloud-based, open IoT operating system”, MindSphere

Industry 4.0 and Industrial Internet of Things 21

Honeywell

EL
 Cloud software service for performance optimization
 Deeper insights of processes, driving agents and design skills
 Efficient solution for oil and gas industries

PT
 Secure, scalable and standards-based platform
 Supports for SaaS business models

Source: “Honeywell Industrial Internet of things-Cloud Software”, Honeywell

Industry 4.0 and Industrial Internet of Things 22

Honeywell (Contd.)

EL
Data analytics

Onsite control and management

Connected World of devices

PT
and assets

Smart and secure alliance

Source: “Honeywell Industrial Internet of things-Cloud Software”, Honeywell

Industry 4.0 and Industrial Internet of Things 23

 References

EL
[1] A. Gilchrist, “Industry 4.0: The Industrial Internet of Things”, Apress, 2016.
[2] A. Ustundag and E. CevikcanIndustry, “4.0: Managing The Digital Transformation”, Springer, 2018.
[3] “NIST Cloud Computing Reference Architecture”, NIST, Available Online:
https://ws680.nist.gov/publication/get_pdf.cfm?pub_id=909505, Accessed on August 20, 2018.
[4] “PREDIX: The application platform for digital industrial solutions”, Predix
Available Online: www.ge.com/digital/sites/default/files/Predix-from-GE-Digital-Overview-Brochure.pdf,

PT
Accessed on August 12, 2018.
[5] “MindSphere The cloud-based, open IoT operating system”, MindSphere,
Available Online: www.siemens.com/content/dam/webassetpool/mam/tag-siemens-
com/smdb/corporate-core/software/mindsphere/siemens-plm-mindsphere-brochure-69167-a19.pdf,
Accessed on August 12, 2018.

Industry 4.0 and Industrial Internet of Things 24

 References

EL
[6] “Honeywell Industrial Internet of things-Cloud Software”, Honeywell,
Available Online: www.honeywell.com/newsroom/news/2016/10/honeywell-launches-cloud-enabled-software-
service-to-optimize-plant-performance, Accessed on August 12, 2018.
[7] “C3 IoT: Products + Services Overview”, C3IoT, Available Online:
idcdigitalsummit.com/COMMONS/ATTACHMENTS/C3IoT_products_services_17_1110.pdf, Accessed on August
12, 2018.

PT
[8] “Predictive Analytics Solutions for Global Industry l Uptake”, Uptake, Available Online: www.uptake.com,
Accessed on August 12, 2018.
[9] “Meshify - Complete IoT Solution”, Meshify, Available Online: meshify.com, Accessed on August 12, 2018.
[10] “The Future of Industrial IoT”, Industrial Internet Consortium, Available Online:
https://www.slideshare.net/M2M_Alliance/the-future-of-industrial-iot-by-stephen-mellor-cto-industrial-
internet-consortium, Accessed on August 20, 2018.

Industry 4.0 and Industrial Internet of Things 25

References

EL
[11] “Will There Be A Dominant IIoT Cloud Platform?”, Element 14, Available Online: https://fr.farnell.com/will-
there-be-a-dominant-iiot-cloud-platform, Accessed on August 12, 2018.
[12] J. Gubbi, R. Buyya, S. Marusic, and M. Palaniswami, “Internet of Things (IoT): A vision, architectural
elements, and future directions”, Future Generation Computer Systems, vol. 29, pp. 1645-1660, 2013.

PT
Industry 4.0 and Industrial Internet of Things 26
 EL
PT
Industry 4.0 and Industrial Internet of Things 27
 EL
IIoT Analytics and Data Management:
Cloud Computing in IIoT – Part 2

PT Dr. Sudip Misra

Professor
Department of Computer Science and Engineering
Indian Institute of Technology Kharagpur
Email: [email protected]
Website: http://cse.iitkgp.ac.in/~smisra/
Research Lab: cse.iitkgp.ac.in/~smisra/swan/

Industry 4.0 and Industrial Internet of Things 1

Industrial Cloud Platforms Providers: Our Discussion

EL
 By industrial companies
 GE Predix
 Siemens MindSphere
 Honeywell

PT
 By Software development firms
 C3 IoT
 Uptake
 Meshify
Source: “Will There Be A Dominant IIoT Cloud Platform?”, Element 14

Industry 4.0 and Industrial Internet of Things 2

C3 IoT

EL
 Platform offers services including analysis and prediction
 Secure framework: authentication and authorization
 Artificial Intelligence powered analytical tools

PT
 C3 Data Lake: Storage service for unstructured data in RESTful
format

Source: “C3IoT: Products + Services Overview”

Industry 4.0 and Industrial Internet of Things 3

C3 IoT (Contd.)

EL
 C3 IoT Platform tools  C3 IoT SaaS Products
 Data Integrator  Predictive Maintenance
 IDE  Inventory Optimization
 Data Explorer  Supply Network

PT
 Analytics Designer  Energy Management
 Ex Machina  Fraud Detection
 Data Science Notebook  Sensor Health
 Type Designer
Source: “C3IoT: Products + Services Overview”

Industry 4.0 and Industrial Internet of Things 4

Uptake

EL
 Enterprise solutions equipped with latest technologies to
provide high value low cost
 Identifying the strength and goals of business through trade
discussions

PT
 Smarter ways of achieving the goal

Source: “Predictive Analytics Solutions for Global Industry l Uptake”, Uptake Digital

Industry 4.0 and Industrial Internet of Things 5

Uptake (Contd.)

EL
Better resource utilization

• Effective cost computations

• Avoid replications
• Growth in production

Technological supports

• Build constant revenue flow for subscribed user services

PT
• Automation and technological advancements
• Ease of buy and sell process

Customer satisfaction

• Secure services
• Smarter storage solutions
• Compliance with legal rules and regulations

Source: “Predictive Analytics Solutions for Global Industry l Uptake”, Uptake Digital

Industry 4.0 and Industrial Internet of Things 6

Meshify

EL
 Provides industrial IoT platform
 Faster development and deployment processes
 Real-time monitoring
 Low-cost solutions

 Now
PT
Solutions:

 Tracker
 Enterprise
Source: “Meshify - Complete IoT Solution”, Meshify

Industry 4.0 and Industrial Internet of Things 7

Meshify (Contd.)

EL
Asset tracking in real-time
with geo-localized data
Now

Fast deployment, real-time

monitoring and visualization

PT
Full stack IoT solution, flexible configuration
support and custom notification facility

Meshify
Tracker
Enterprise

Source: “Meshify - Complete IoT Solution”, Meshify

Industry 4.0 and Industrial Internet of Things 8

Cloud-Platform for Device Management

EL
 Need for device management:
 Increase in number of devices makes an IIoT ecosystem more complex
 Not deploy and forget scenario for installed devices
 Change in standards and services

PT
 Replacement of faulty devices
 Security requirement
 Device management is dependent on few other functionalities
 Better way to keep device management service at cloud
Source: “Fundamentals of IoT device management”, IoT Design

Industry 4.0 and Industrial Internet of Things 9

Cloud-Platform for Device Management

EL
 Features that cloud platform provider should offer for device
management:
 Provisioning and authentication
 Fault diagnosis and monitoring


 PT
Updates, security patches and maintenance
Configuration and control
Device decommission

Source: “Fundamentals of IoT device management”, IoT Design

Industry 4.0 and Industrial Internet of Things 10

Cloud-Platform for Device Management

EL
 Example of cloud platform providers with device management
facility:
 Bosch IoT Remote Manager
 AWS’s IoT Device Management


 PT
Verismic’s Cloud Management Suite
ICP DAS’s IoTstar
Software AG’s Cumulocity

Industry 4.0 and Industrial Internet of Things 11

Service Level Agreement (SLA) for IIoT

EL
 Many IIoT applications are real-time and include safety
measures
 Framework should achieve the goals as per plan
 Services should be as per the agreement with cloud provider

PT
 A SLA helps the cloud provider in promising the deliverables
 SLA helps the industrial client to check what and how good
the cloud provider gives service
Source: Papadopoulos et al. , 2017

Industry 4.0 and Industrial Internet of Things 12

Characteristics of a Good SLA

EL
Affordable Achievable

SLA
Mutually Acceptable Characteristics Meaningful

PT
Controllable

Source: Sturm et al., 2000

Quantfiable

Industry 4.0 and Industrial Internet of Things 13

Current Status and Future of SLAs in IIoT

EL
 SLA complied cloud service in IIoT is at infant stage for following
reasons:
 Quality of services offered has interdependencies
 Methodologies and frameworks of IIoT are not well developed

PT
 Life cycle management of an SLA in industrial context is not clear
 Lack of SLA enforcement policies for both provider and consumer
 SLA support for IIoT is crucial along with business models
 Future IIoT needs a standardization of SLA and its management
Source: Papadopoulos et al. , 2017

Industry 4.0 and Industrial Internet of Things 14

Choosing the Right Cloud Vendor for IIoT

EL
 Cloud is the heart of an IIoT ecosystem and choosing the
correct platform is crucial
 Market of Many cloud vendors available with similar services
 A proper checklist of needs and cross checking with services

PT
from vendors

Source: “Top 10 selection criteria to choose your IoT platform”, IOTIFY

Industry 4.0 and Industrial Internet of Things 15

Choosing the Right Cloud Vendor for IIoT

EL
 Points to consider:
 Scalability support
 Bandwidth requirement
 Communication protocols




PT
Security
Interoperability
Edge Intelligence feature
Infrastructure management
Source: “Top 10 selection criteria to choose your IoT platform”, IOTIFY

Industry 4.0 and Industrial Internet of Things 16

Limitations in Cloud-Based Approach

EL
 Volume, velocity and variety
 Higher latency
 Bandwidth requirement for huge data volume

PT
 Reliability for the big network
 Need for scalable security

Source: “Introduction to Edge Computing in IIoT”, Industrial Internet Consortium

Industry 4.0 and Industrial Internet of Things 17

Centralized vs. Decentralized Approach

EL
 Cloud based centralized approach suffers from many
limitations
 Decentralized approach decreases the load at cloud
 Real-time operations feasibility

PT
 More scalable IIoT network and features
 Greater mobility support

Source: “Today's Centralized Cloud And The Emerging Decentralized Edge”, Forbes

Industry 4.0 and Industrial Internet of Things 18

Industry 4.0 Objectives

EL
 Robust solutions
 Higher production
 Better Customer satisfaction

PT
 Expanded security
 Better performance
 Entire world of industry at one place

Source: “Industry 4.0: The Industrial Internet of Things”, Apress, 2016

Industry 4.0 and Industrial Internet of Things 19

Industry 4.0 Requirements from IIoT

EL
 Aims to achieve greater production, optimized decisions,
efficiency and availability
 Deeper insights of analysis and prediction
 Establishing a connected world of machines, systems,

PT
products and environments
 Collection of data from each sector and performing analytics
to exploit the wealth at its best
Source: “Industry 4.0: The Industrial Internet of Things”, Apress, 2016

Industry 4.0 and Industrial Internet of Things 20

Solution

EL
 A Decentralized or distributed approach along with cloud
 Handling time-sensitive data
 Immediate action and quick response

PT
 Delay in proper action at proper time may create hazardous
situation
 Thus, Fog emerges to be a solution

Source: “Introduction to Edge Computing in IIoT”, Industrial Internet Consortium

Industry 4.0 and Industrial Internet of Things 21

Fog Computing

EL
 An added layer between the edge and the cloud layer
 Not a replacement but an addition to cloud
 Identify useful data thus reducing the amount of raw data

PT
sent to cloud
 Increased scalability with reduced traffic

Source: Bonomi et al., 2014

Industry 4.0 and Industrial Internet of Things 22

Fog Computing (Contd.)

EL
 Intelligent devices deployed at edge
 Intelligent compute devices known as fog nodes
 Intelligent in providing services like filtering, aggregation and

PT
translation
 Distributed at one level, centralized on the other

Source: Bonomi et al., 2014

Industry 4.0 and Industrial Internet of Things 23

References

EL
[1] A. Gilchrist, “Industry 4.0: The Industrial Internet of Things”, Apress, 2016.
[2] A. Ustundag and E. CevikcanIndustry, “4.0: Managing The Digital Transformation”, Springer, 2018.
[3] S.Sarkar, S.Chaterjee, and S.Misra, “Assessment of the Suitability of Fog Computing in the Context of
Internet of Things”, IEEE Transactions on Cloud Computing, vol. 6, no. 1, pp. 46-59, 2018.
[4] A. Mohammad, S. Zeadally, and K. A. Harras, “Deploying Fog Computing in Industrial Internet of Things
and Industry 4.0”, IEEE Transactions on Industrial Informatics, 2018. DOI: 10.1109/TII.2018.2855198.

PT
[5] “Will There Be A Dominant IIoT Cloud Platform?”, Element 14, Available Online: https://fr.farnell.com/will-
there-be-a-dominant-iiot-cloud-platform, Accessed on August 12, 2018.
[6] “C3 IoT: Products + Services Overview”, C3IoT, Available Online:
idcdigitalsummit.com/COMMONS/ATTACHMENTS/C3IoT_products_services_17_1110.pdf, Accessed on
August 12, 2018.
[7] “Predictive Analytics Solutions for Global Industry l Uptake”, Uptake, Available Online: www.uptake.com,
Accessed on August 12, 2018.

Industry 4.0 and Industrial Internet of Things 24

References

EL
[8] “Meshify - Complete IoT Solution”, Meshify, Available Online: meshify.com, Accessed on August 12, 2018.
[9] “Top 10 selection criteria to choose your IoT platform”, IOTIFY, Available Online: https://iotify.io/top-10-
selection-criteria-for-your-iot-cloud-platform/, Accessed on August 23, 2018.
[10] A. V. Papadopoulos, S. A. Asadollah, M. Ashjaei, S. Mubeen, H. Pei-Breivold, and M. Behnam, “SLAs for
Industrial IoT: Mind the Gap”, In proc. of 5th International Conference on Future Internet of Things and
Cloud Workshops, 2017.

PT
[11] S. Mubeen, S. A. Asadollah, A. V. Papadopoulos, M. Ashjaei, , H. Pei-Breivold, and M. Behnam,
“Management of Service Level Agreements for Cloud Services in IoT: A Systematic Mapping Study”, IEEE
Access, vol. 6, pp. 30184-30207, 2017.
[12] “Introduction to Edge Computing in IIoT”, Industrial Internet Consortium, Available Online:
www.iiconsortium.org/pdf/Introduction_to_Edge_Computing_in_IIoT_2018-06-18.pdf, Accessed on
August 23, 2018.

Industry 4.0 and Industrial Internet of Things 25

References

EL
[13] “Today's Centralized Cloud And The Emerging Decentralized Edge”, Forbes, Available Online:
https://www.forbes.com/sites/forbestechcouncil/2017/12/05/todays-centralized-cloud-and-the-
emerging-decentralized-edge/#3bc981c46b3c, Accessed on August 23, 2018.
[14] F. Bonomi, R. Milito, P. Natarajan and J. Zhu, “Fog Computing: A Platform for Internet of Things and
Analytics”, Springer, 2014.
[15] “Fundamentals of IoT device management”, IoT Design, Available Online: http://iotdesign.embedded-

PT
computing.com/articles/fundamentals-of-iot-device-management/, Accessed on August 23, 2018.
[16] R. Sturm, W. Morris, and M. Jander, “Foundations of Service Level Management”, ser. Sams Professional
Series. SAMS, USA, 2000.

Industry 4.0 and Industrial Internet of Things 26

 EL
PT
Industry 4.0 and Industrial Internet of Things 27