Bringing the power of IBM Watson IoT to the Edge with Cisco Dave Locke Senior Inventor IBM IoT Ecosystem Manager Connecting Things that Matter @DaveJLocke
Bringing the power of IBM
Watson IoT to the Edge with
Cisco
Dave Locke Senior Inventor IBM IoT Ecosystem Manager Connecting Things that Matter @DaveJLocke
Agenda
The Power of Data
IBM Watson IoT Platform
Edge processing
Use Cases
Summary
3
4
IoT is Driving Digital Disruption Into the Physical World
Advanced Analytics
Product Lifecycle Mgmt
Cloud Computing
Pervasive Connectivity
Embedded Sensors
Creating New Products and Business Models
Smarter, safer cars
Health and fitness
Home and building
automation
Improving Operations and Lowering Costs
Predictive maintenance
Analyze and reduce risk
Factory automation
Driving Engagement and Customer Experience
Smarter, more profitable retail
Engaged events and venues
Apps that link the digital and
physical world around a brand
Accelerating advancements in technology… Are transforming every part of business…
Leveraging the data generated by digital technology provides intelligence to help us do things better, improving our responsiveness and ability to predict and optimize for future events
INTELLIGENT
Digital technologies (sensors and other monitoring devices) are being embedded into many objects, systems and processes
INSTRUMENTED
INTERCONNECTED
In the globalized, networked world, people, systems, objects and processes are connected, and they are communicating with one another in entirely new ways
Smarter Planet and the Internet of Things
Little Data Big Data
IoT Driving Forces…
Price Power conservation,
Energy Generation
Form Factor, Miniaturization
Connectivity,
Network
Drive Innovation Edge
Most IoT data are not used
currently. For example, only
1 percent of data from an oil
rig with 30,000 sensors is
examined. The data that
are used today are mostly
for anomaly detection and
control, not optimization and
prediction, which provide
the greatest value.
Analytics
Data • Cloud • Big Data • Analytics • Applications
IoT Device
Data
Traditional Data Processing Model
Traditional: Deliver Data to the Analytics
9
Anatomy of an IoT Solution
9
Watson IoT Platform
Sensors, Devices,
Gateways & Networks
Other
Data Sources
Weather
Map
01 0110 0010 001001
Devices Platform Applications
Other IoT platforms BMS
Asset
IBM Watson IoT Platform
10
11
Third Party Apps Offerings
IBM Watson IoT Platform Connect Attach, Collect & Organize, Device Management, Secure
Connectivity, Visualization
IBM Watson IoT Platform Information
Management Storage & Archive, Metadata Management, Reporting, Streaming
data, Parsing and Transformation, Manage unstructured data
IBM Watson IoT Platform Analytics Predictive, Cognitive, Real-time, and Edge
IBM Watson IoT Platform Risk Management Security Analytics, Data Protection, Auditing/Logging,
Firmware Updates, Key/Cert Mgmt, Org Specific Security
Third Party Apps
The IBM Watson IoT Platform Everything you need to Innovate with IoT
13
IoT requires the right capabilities applied to the right data for the right
results
Cognitive technology
enables deeper customer
engagement through
enhanced interactions and
automated discovery and
insights using machine
learning techniques
Predictive models are
created from historical data
to generate insight and
recommend actions before
situations cause business
disruptions
Real-Time analytics enables
monitoring and processing of
streaming data to enable
“perishable insights” and
automated decisions in near
real-time
Real-Time
Most machine data is
worthless about 1
second after it is
generated
Cognitive
IoT will rapidly
change our ability to
interact with
machines and
engage customers
Predictive
70% of the most
profitable
companies will
leverage predictive
analytics in 2016
In the cloud & at the edge
15
Real-Time Watson IoT Platform Analytics Real-Time Insights
15
Real-time
dashboard
Recommendations drive response in Maximo
5
Sensors provide information about the device
1
Device
SCADA, historians
Data may be collected by a gateway device for connectivity or protocol translation
IoT Connect
IoT Analytics
Real-time data
Rules trigger an action, such as an alert, email, text message or a work order in Maximo
4
Data drives real-time analytics and business rules
3 Data comes in through Watson IoT Platform Connect
2
Maximo
1
2
• Contextualizes device data
• Monitors streaming data to detect situations
• Acts on insights from the data
Data is enriched with external data such as Weather or asset master data
1 3a
16
ingest
analyze report & recommend act
profile SaaS offering
Prebuilt analytics
Faster time to value
Designed for line of business
Reduces need for data scientists
Insight at point of engagement
Predictive IBM Predictive Maintenance on Cloud
IBM Predictive Quality on Cloud
IBM Predictive Warranty on Cloud
16
17
Watson IoT API families allow easy integration of
cognitive analytics into IoT apps
Natural Language Processing
Enables interaction through natural human
language and dialog
Machine Learning
Automates data processing and continuously
monitors new data to learn and improve results
Textual Analytics
Enables mining of textual sources to find
correlations and patterns in these vast amounts
of untapped data
Video/Image Analytics
Enables monitoring of unstructured data from
video feeds and image snapshots to identify
scenes and patterns
Analytics at the Edge
18
Why This is So Unique
Traditional: Deliver Data to the Analytics
Analytics
Data
Edge Node
Fog Node
IoT Device
Analytics Analytics Analytics
Analyze Data in the 'Right' Place by Distributing Analytics from Cloud to Edge
Data Data Data
• Cloud • Big Data • Analytics • Applications
This is a Differentiated Route from the Industry Direction
IoT Device
Data
© 2016 Cisco and/or its affiliates. All rights reserved. Cisco Public
Combined Architecture
Enable Cognitive Computing Enable edge analytics; route to the cloud
&
Cloud Edge Node Fog Node IoT Device
Processing Processing Processing
Data Data Data
21
Analytics at the Edge
Define and manage analytics in the cloud, run them where it makes sense
Analyze & act on data close to source
Reduce burden on constrained networks and reduce transmission costs
Enable continuous operations even if the network is down
Deliver high value data to the cloud for richer cross site / cross fleet analytics
IBM and CISCO announce analytics from Cloud to Edge!
Edge
22
Expanding Analytics Further Into the IoT Environment
IBM Watson IoT Platform
(Cloud)
Capabilities:
• Complex analytics
• Analytic definition and distribution to
edge
• Longer term trends
• Pattern detection and machine
learning
Edge Gateway
WIoTP Edge
Analytics Agent
Operations:
• Filter and reduce data sent to cloud
• Pre-process and transform raw
data
• Identification of critical conditions to
send to cloud for additional
analytics
• Drive actions as the result of
analytics
IBM IoT Platform Analytics: An integrated
cloud-and-edge analytics programming model
that allows control and optimization over the
data flowing between edge and cloud.
IoT Devices
23
Where Does the EAA Run?
IBM Watson IoT
Platform
IoT Devices IoT Gateway
• EAA will run on IoT Gateway devices made by companies we
partner with
• IBM is partnering with Cisco today and will be partnering with
other gateway providers in the future
Edge ….
24
Combine with Cloud Analytics For Added Value
IoT
Gateway
with EAA
IBM Watson
IoT Platform
Edge
Analytic
Results
Service
Request
Device
Commands
• Analytics on the edge can send
data resulting from analytics to
the cloud for additional analytics
• Can be combined additional IoT
data for additional and analytics
and drive cloud based actions
24
25
How it works
CLOUD ON-PREM
IoT Device IoT Device
IBM Cognitive Analytics Agent
Broker; Cisco Edge,
Fog Computing
& Edge Analytics
Watson IoT Platform
IBM Real-time Insights
Gateway Deploy to
IBM Edge
Engine (EAA)
2
3
Device Data
Flows into
the Edge
4
IBM EAA Filters &
Aggregates Device
Data, Rules Trigger,
Drive Alerts & Actions
Local Actions
Go Back Out
to IoT Devices
4b
Data, Alerts, &
Cloud Actions Flow
Back to Cloud
5
Enrich with Context
(Weather) & Apply
Deeper Cognitive,
Predictive Analytics
4a
1 Configure Rules &
Actions in the Cloud
Actions
Analytics
26
A progression of analytics & capabilities at the edge…
Real-Time High speed, ‘perishable’ data require
scalable contextualization and processing
to gain insight and react in near real time
Complex systems need more natural
interaction patterns via voice and chat
that operate independently of the cloud
Natural Language
Processing
Data is filled with trends, such as rising
temperature or cyclical patterns in a
motor’s RPMs, we need to automatically
understand norms and forecast issues
Machine Learning
Edge Workflows & Transactions Increasingly, we’ll need to handle more complex
logic and transactions at the edge, extending
insights to more complex orchestrations of
actions with enhanced security via blockchain
Unstructured data is also proliferating in the form of
video, image &audio data. This data needs to be
correlated with other sources of machine data and
processed for insights
Unstructured data
Predictive Mission critical equipment and
processes need to run smoothly, and
you need advance warning of issues in
order to avoid down time, business
disruption, and safety issues
Use Cases
27
28
Targeting three operational patterns
Autonomous Operations
Remote Operations
Large Scale Operations &
Fleets
Industries: Automotive, Oil & Gas, Manufacturing, Heavy Equipment
Examples: Discrete manufacturing & continuous operations
Potential: Economic impact of $1.2 trillion to $3.7 trillion per year in 2025 (McKinsey)
Benefits
• 10 – 20% reduction in health & safety costs
• 5 – 10% increase in worksite productivity
• 5 – 10% reduce costs of equipment
Industries: Commercial Real Estate, Travel & Transportation, A&D, Heavy Equipment, Electronics
Examples: Elevators, motors, aircraft & engines, buildings & systems, commercial equipment
Potential: Economic impact of $560 billion to $850 billion per year by 2025 (McKinsey)
Benefits
• 10 – 20% increase in personnel productivity
• 5 – 12.5% decrease in logistics & scheduling costs
• 10 – 40% cost savings for equipment & maintenance
Industries: Transportation, Oil & Gas, Utilities, Mining, Construction
Examples: Shipping, Drilling, pipelines, oil platforms, wind/solar farms
Potential: Direct economic impact of $160 billion to $930 billion per year in 2025 (McKinsey)
Benefits
• 10 – 20% increase in productivity
• 5 – 12.5% decrease in operation costs
• 10 – 40% cost savings for equipment & maintenance
Global Auto Manufacturer benefits from edge-based Condition
Monitoring & Predictive Maintenance
Challenges
• Ensure high quality welds made by robots during manufacturing, improve detection speed to reduce impact of down process activities
• Monitor robot health through predictive modeling to detect early signs of deteriorating performance and risk of failure
Solution
• Edge analytics for real-time monitoring of welding robots based on properties such as vibration, rotation speed, velocity and weld temperature
• Cloud-based cognitive analytics for forecasting asset health and predicting component failures
• Components: IBM Predictive Maintenance & Quality, IBM Watson IoT Platform & Edge Analytics, Cisco Edge Analytics Fabric
Outcomes
• Higher quality welds with reduced rework, overtime, and scrap improving output and decreasing overall costs
• Predictability of robot issues allowing for pro-active maintenance during operational down time
Port of Cartagena leverages Condition Based Maintenance
Challenges
• Fleet of hundreds of vehicles, cranes and boats operating 24x7x365. Struggling to maintain equipment efficiently.
• Can’t afford to rely exclusive on cloud analytics due to potential connectivity problems.
Solution
• Consists of: Cisco UCS240 Server, Cisco Edge Analytics Fabric,
Watson IoT Platform with Edge Analytics.
• Optimizing maintenance by triggering automatic alerts based on
conditions at the edge (fuel levels, battery voltage, engine conditions
and other advanced measures).
Outcomes
• Now conducting condition-based maintenance, informed by actual
condition of assets operating at the edge.
• Critical data analyzed immediately at the edge; high-value data sent
data for deeper analysis in the cloud.
Utility improves outage detection and notification
Challenges
• Equipment failures and storms resulting in outages in the electrical grid
• Gaining real-time understanding of emerging situations to respond quickly and appropriately
• Notifying customers of the issue, current status and estimated restoration
Solution • Consists of: Cisco router (pole-top mounted), Cisco Edge Analytics Fabric, Watson IoT
Platform with Edge Analytics
• Help utility identify power outages faster by bringing analytics to the edge of grid to monitor smart meter telemetry and pinpoint outages as they occur
• Forward analytic results back to the cloud for more powerful analytics, wide-area intelligence & cognitive learning
Outcomes
• Improved notification time for the utility
• Lower operating expenses, Increased customer satisfaction
• Improved awareness and faster, proactive decisions through improved analytics
Silverhook Powerboats
Challenges
• Operating engines that are costly and dangerous to damage; rely on engine governors which adversely impact performance.
• Need to monitor real-time engine conditions and get feedback to operator with low latency.
Solution
• Consists of: Cisco IR829 Ruggedized Network Router, Cisco UCS240 Rack Server, IBM Watson IoT Edge Analytics
• Created rules at the edge, triggering alarm based on engine condition. New dashboard shows real-time race data, including status, engine condition, speed, RPMs and more.
Outcomes
• Better real-time monitoring with low latency.
• Helps Silverhook push for maximum performance with confidence,
enabling them to avoid engine shutdowns and win more races.
Silverhook Powerboats
Summary
34
35
Summary
• Analytics are a key to gaining insights from IoT data
• Scalable solutions require a variety of analytics performed on the right data…and at the right location—including real-time, predictive, and cognitive & performed from edge to cloud
• Cognitive analytics will enable us to deliver transformative solutions that interact with users naturally and can learn from and automatically process the flood of data
• IBM has the portfolio of analytics to help customers succeed with IoT solutions and a network of partners to help deliver
35
36
Learn more about IBM’s point of view on the Internet of Things ibm.com/IoT
Try out Internet of Things on Bluemix ibm.biz/try_iot
Try out Real-Time Insights
ibm.biz/try_rti Try out Edge Analytics
https://ibm.biz/Bdsdzs Getting Started video for Real-Time Insights
youtu.be/_Q4GlqAf2m4 Join us in our IoT conversations
@IBMIoT
IBM IoT – Get started today
36
37
38
Edge Applications
Condition Based Maintenance
Condition Based Maintenance (CBM) uses sensor data from equipment and applies a monitoring strategy that uses the actual condition of the asset to decide when and what maintenance should be done. CMB can augment a time-based maintenance strategy and helps reduce failures while reducing the cost of maintenance overall by right sizing maintenance intervals. Operations benefits from greater asset availability and better predictability of performance.
Predictive Maintenance
Predictive Maintenance applies a deeper analysis of historical data to build predictive models for asset health and failures. Predictive models are then used to give forewarning of failures giving operations and maintenance the time to address impending issues with decreased risk of failure. Predictive models can be developed as an extension of CBM and used to understand potential failures of equipment in real-time.
Predictive Quality
Predictive Quality works holistically across equipment and work cells to understand the predictors of poor quality across a process. Predictive Quality applies statistical modeling to historical data from across equipment to generate predictive quality models for the entire process. Predictive Quality uses data such as environment or weather conditions and asset properties.