2. The 4th Industrial Revolution
18th Century
Industry 1.0
Mechanical
production powered
by water and steam
20th Century
Industry 2.0
Mass production based
on the division of labour
and powered by
electrical energy
70s
Industry 3.0
Electronics and IT for
a further automation
of production
Today
Industry 4.0
Cyber physical
production systems
+
Technology
Progress
Smart
Devices
Source: McKinsey & Company
3. The expected benefits of Industry 4.0
Flexibility
Speed
Productivity
Quality
Competitiveness
Increased Flexibility thru production of small lots at large
scale costs
Increased Speed from prototype to series production thru
innovative technologies
Increased Productivity thru reduced set up times, error
reduction and machine downtime
Increased Quality and reduced waste parts thru sensors
that monitor the production in real time
Increased product Competitiveness thru broader
functionalities triggered by Internet of Things
4. Digitization as an opportunity
Procurement /
Development
Manufacturing
Marketing, Sales,
Service/After Sales
AWS for lifting &
shifting existing
applications
• CAD/HPC
• PLM/Collab
• ERP
AWS for cloud
optimized
workloads:
• Big Data
• Industrial IoT
• Analytics
AWS to
• Intensify the relationship
to your customer
• Leverage on Partner
Ecosystem Support
5. What are the technical enablers of Industry 4.0 ?
Autonomous
robots
Simulation
Horizontal and vertical
system integration
The Industrial
Internet of Things
Cybersecurity
Cloud
Additive
manufacturing
Augmented
reality
Big Data
and Analytics
Source: McKinsey Company, Boston Consulting Group
Cloud is one of 9 stand-alone enablers
6. What are the technical enablers of Industry 4.0 ?
Autonomous
robots
Simulation
Horizontal and vertical
system integration
The Industrial
Internet of Things
Cybersecurity
Cloud
Additive
manufacturing
Augmented
reality
Big Data
and Analytics
Cloud is the central innovation enabler across all disciplines and industries
9. Connecting devices to cloud applications
requires undifferentiated heavy lifting.
Many SDKs
& Tools
Alternate
Protocols
Scalability Security &
Management
Integration with Cloud
and Mobile Applications
10. Introducing AWS IoT
Respond to signals from your
fleet of devices and take
action with Rule Engine
Connect any device via
MQTT/HTTP securely. Quickly get
started with AWS IoT Starter Kits
and Scale to billions of messages
across millions of devices
Securely connect any
physical device to AWS
Shift business logic from
device to cloud and route data
to AWS service of your choice
for storage and analysis using
rules engine.
Create Web and Mobile
Applications that Interact with
Devices reliably at any time
“Securely connect one or one-billion devices to AWS,
so they can interact with applications and other devices”
Easily build applications on
web and mobile that interact
with devices, even when they
are offline, with AWS SDK and
Device Shadow.
11. AWS IoT Platform
All in one service
Message Broker
+ Rules Engine
+ Shadow
+ Registry
All for $5/M Msg*
Managed service
No installation
Automatic scaling
No pre-provisioning
Redundant across AZ
Pay as you go
* Varies by Region
12. Amazon Machine Learning
> Predict() Function
SELECT predict(model ID) as prediction FROM /device/data WHERE temperature > 150
ACTION trigger a Lambda function
Three types of Prediction
1. Binary Classification (one of two
possible classes)
2. Multiple classification (one of more
than 2 outcomes)
3. Regression (predict a numeric
value)
Model training
Posting the data to S3 will trigger a
new training and make the model
better
13. Anomaly Detection - Predictive MaintenanceAnomaly Detection – Predictive Maintenance
You start by training a machine learning model with data that corresponds to normal expectations. This way, new data, which
does not correspond to the normal expected values will be detected instantly. This way you can bring a confitiong monitoring
system to automatically learn from the regularly performed diagnosis.
Upload data and train ML model
Amazon
S3
Amazon
Redshift
Amazon
Machine
Learning
Detect anomalies and send out warnings
AWS IoT Amazon
CloudWatch
Email Messaging
Amazon
SNS
Amazon
Machine
Learning
PLC
14. The Customer: a Leading Supplier in Automotive
Manufacturer of Highly technological
components
10.000 employees around the World
Located in 14 Countries
Sales in more then 100 countries
Revenues of $2.7 billion
Industrial & Transportation
15. The Challenge: Reducing Waste
The Lean “Kaizen” (change good)
enabled by AWS IoT and AWS ML
Reduce the number of fouled
components
Make the production line more and more
intelligent
Continuous improvement of factory
settings to avoid wasteChange Good
16. The Solution: IoT Prediction Platform
IoT Prediction platform is composed of:
A Machine Learning model able to process
recipes data
List of possible data correlation
Leverage one AWS IoT and AWS Machine
Learning Architecture in order to ingest and
process all incoming data
Model Configuration
17. Serverless architecture is state of the art paradigm offering great advantages:
No servers; Low maintenance and administration effort
No single point of failure - HA and Reliability
Highly scalable and elastic
No fixed costs; Pay-per-use resources
No lock-in; Standard protocols used
Security based on standard AWS services
DevOps approach with CI, CD & AD
Global reach easily replicable using AWS regions
Ready to support future device types
The Solution: Serverless is better
19. Elasticsearch Integration
> Elasticsearch Action
Enable visualization
Leverage Kibana for fast and easy
visualization of data
Enables complex search queries
Averages, time bound, and more
21. Smart Manufacturing: customer’s needs
Enterprise systems to
Manufacturing systems
integration
Manufacturing KPI
evaluation and
optimization
Big data analysis and
machine learning for KPI
prediction
Global visibility of
manufacturing plants
performance
McKinsey Global Institute Analysis
23. THINK BIG
IoT drives the digital
transformation of all
things
Connect things to IT
and to business
processes to achieve
business outcome
START SMALL
Select a viable use
case with fast return
on investment
Proof of Concept
SCALE FAST
Learn from early
insights and automate
as you expand
Leverage Cloud to
scale fast
Lessons learned
26. Business Benefits
Increase visibility on manufacturing facilities worldwide
Define a unique semantic layer for manufacturing KPIs
Use Kibana self service composition environment for tailor
made dashboarding
Access to “Decreto Industry 4.0” fiscal benefits for Italian
plants