This presentation was given to an executive MBA session at UCSD in April 2016. The session reviewed big data, internet of things, and how companies are gaining value from location, sensor, manufacturing and other data to make better business decisions.
3. 3
Data Growth
Source: IDC - sponsored by EMC. As the Economy Contracts, the Digital Universe Expands. May 2009
Transactions
Interactions
1024
1021
1018
1015
1012
109
Yottabyte
Zettabyte
Exabyte
Petabyte
Terabyte
Gigabyte
4. 4
IoT, Sensors, and Tiny Computers
Internet of Things
SensorsEmbedded
computers
Industry Specific
(eg CatScans, etc.)
Data Center
systems
5. 5
• Events generating data
– Vibration
– Temperature, humidity
– Wind speed, direction
– Air/liquid flow or pressure
– Location, navigation
– Tilt level, rotation
– Light, sound
– Radiation, chemicals
– Biological
- Heart rate, blood pressure
- Brain activity, chemicals
– Inventory, sales (RFID)
• Data format: JSON or proprietary
The Data Sensors Collect
6. 6
• A new wave of devices will help you track health statistics but security and
privacy concerns loom for this health “Big Data”
Google Glass Jawbone UP Ingestible
Sensors
The Quantified Self
7. 7
Gartner: Growth of the Internet of Things
Source: Gartner, Forecast: The Internet of Things, Worldwide, 2013, Nov 2013
Billions of Things in Use
Connected PC, smartphone, tablet IoT
0
5
10
15
20
25
30
2009
2020
13. 13
New Tech for Vehicles Focused on Safety & Autonomy
Carmakers are facing seismic change. Suppliers which
were largely kept under the hood are set to grow in
influence as the industry adds more and autonomous
features to vehicles
Detects close range
objects to aid parking and
avoid collision by using
radio waves
Integrates driver
assistance functions:
algorithms for every
scenario
Semiconductors underpin
advanced electronic
functionality
Seeks longer range objects
for use in Adaptive Cruise
Control systems
Used in front and rear
parking sensors in modern
cars. Will be adapted for
assisted parking and short
range pedestrian/obstacle
detection
Enables in-car night vision
systems that can detect
objects further away than
traditional headlights to
avoid collisions at night
Integrates both directional and
distance information in lane
departure systems and traffic sign
recognition
For precise navigation
Allows vehicles to
communicate with
each other
Advanced driver assistance systems
Ultrasound
Front/rear
short radar
Infrared
Vehicle to
vehicle
communication
Advance
mapping
Semi
conductors
Long range
radar
Stereo
cameras
Software
14. 14
• Customer centric experience –
connected cars and hazards ahead
• Dealer Analytics –Scorecards and
Action Plan
• Warranty/ Repair/ Failure
Predictions
• Add New Services
• Featured Used/Not Used- learn for
the future
• Buy and customize online – most of
sales activity is done prior to walk-
in’s… “Gamechanging”
Connected car Driver Model
Case Study: IoT at Volvo
Not all sensors or micro computers are part of the internet of things. IoT mainly points to machines and devices that communicate over the internet via Wifi, 3G, 4G, or Ethernet expecting some receiving computer to use the data and provide assistance. But let’s start with the keyword here: THINGS. IoT does not include all the computer servers in the world. We already have that – its called a data center or a Local Area Network, or a home office network. The IoT name describes the trillions of THINGS that are not full computers but are emitting data.
Also notice that sensors range from simple electronic devices to small embedded computers. The embedded computers are constantly growing in capacity due to Moore’s law and economics. So the sensor that detects the refrigerator door opening someday evolves to a micro computer keeping track of inventory inside the refrigerator. Similarly, the on board computer in a Ford Focus evolves to collecting data on driving habits, brake wear, and other maintenance or performance objectives. This can then be used within customer service, engineering and design or warranty to improve insights in these areas.
The business question is how valuable is what specific data? There is a lot of redundant data - is all of it needed? If not, then how to determine what is valuable. Additionally, as sensors increase in computational power and calculations are pushed to the sensors, then derived data may become the most valuable.
Sensors are most frequently used for MEASUREMENT. Types of measurements include: acceleration, tilt, shock, rotation, temperature / humidity / atmospheric conditions, pressure, vibration, strain, force, location, time (traffic movement, usage (minutes/hours of run, dwell, idle time)), engine data (e.g. speed/RPM), miles per gallon, component performance, diagnostic codes, location and status tracking, and more.
Data transmission timing varies by sensor and network availability. It may be in real time, or it may be in batch. For example, many of the early onboard diagnostic sensors in automotive are uploaded in batch once they reach a dealership vs streaming.
Sensors collect all kinds of data. JSON is one of the more common data formats because its small and easy to implement. However, many manufacturers of sensors emit proprietary data formats along with an API programming manual. Its not hard to work with the proprietary data but its also not as easy as selecting an icon in Informatica’s PowerCenter or in DataStage.
If the data format is JSON, use Teradata 15.0 or an ETL tool to easily consume the data.
The growth in IoT will far exceed that of other connected devices. By 2020, the emergence of mass market smartphones and tablets, combined with the mature PC market, will result in an installed base of about 7.3 billion units, which compares with the expected human population of 7.7 billion in that year (based on information from the United Nations Population Division). In contrast, the IoT will have expanded at a much faster rate, resulting in an installed base of about 26 billion units at that time. Installed base is important because it drives the value of service revenue, aggregate communications bandwidth and data center activity. Due to the low cost of adding IoT capability to consumer products, we expect that ghost devices with unused connectivity will be common.
In addition, enterprises will make extensive use of IoT technology, and there will be a wide range of products sold into various markets, such as:
Advanced medical devices, including surgical tools, instrumentation and wearable medical sensors/monitors, and ingestible devices, such as smart pills
Factory automation sensors and applications in industrial robotics
Sensor motes for increased agricultural yield
Automotive sensors
Infrastructure integrity monitoring systems for diverse areas, such as road and railway transportation, water distribution, and electrical transmission
The most popular IoT sensors initially installed are those on items that provide remote monitoring (eg. wind turbines, ATMs/ kiosks, heavy equipment, drilling equipment, etc.) or are mobile (autos, aircraft & engines, people, packages, rail, medical equipment, etc.). This will evolve quickly.
I want you to take away a VERY SIMPLE view of what Internet of Things is. So here is a very simple diagram that will clarify it.
The Internet of things is an expansion of our notion of the internet, to include things….large and small.
And the things are able to, through sensors and tiny chips embedded in them, communicate their state.
And they do this through a series of networking devices called Gateways, which aggregate and package up the data for sending onward, networks themselves, that convey the sensor data either to Public Clouds, or Private Data Centers.
So that companies can do analytics on them.
That’s the simple view.
What is terrifying companies – our customers, is the prospect that this volume of data is yet again another couple of orders of magnitude bigger then when it was just the internet of people and their devices.
The data produced by the IoT is doubling every two years. This creates a "fear" in our customers (budgeting, complexity, executiion, opportunity loss)
What kind of analytics willl be done in the edge? That’s the $64k question.
Event Message (DEMN): DENMs are event-triggered messages broadcasted to alert road users of a hazardous event. Both CAM and DENM messages are delivered to vehicles in a particular geographic region: to the immediate neighborhood in case of CAMs (single hop), and to the area affected by the event for DENMs (multi-hop).
Safety Messages – sent from Vehicle
Probe Data: Data collected from sensors on the car or on the road.. Summarized vs detailed.. There are several papers on how to “summarize” the data through algorithms and modeling
ADAS: Advanced Driver Assist
DMA: Smart phones communicating with vehicle in support of safety, transportation data, smarter travel etc..
Slef-Service – schedule service apporintments, user reset alerts
Usage based services –used car valuation..cross brand usage if own other brands
Telematics tailord services – insurance, parking
In addition to delivering an enhanced level of in-vehicle consumer services, Teradata also has the ability to capture, analyze and report on vehicle performance sensor data to determine potential mechanical failures BEFORE they happen.
This level of predictive maintenance allows Teradata to help car makers improve vehicle safety, maintenance and customer satisfaction.
Today new vehicles are equipped with hundreds iof diagnostic sensors that are equipped to capture data and alert drivers of pending or actual parts/system failures.
These sensors monitor everything from engine performance to breaking, steering, blind spot detection, inclement weather and driver alertness in the vehicle. As these sensors transmit large volumes of vehicle performance data, its important to have an analytics system in place to detect potential failures and determine the size and scope of a potential problem across multiple vehicles or an entire model fleet.
Often a vehicle part failure across multiple vehicles or an entire fleet can trigger a root cause investigation. The ability to quickly and easily capture, analyze and triage a maintenance problem is often the difference in both improving driver/passenger safety and mitigating large scale vehicle outage interruptions.