Introducing
Your new presentation assistant.
Refine, enhance, and tailor your content, source relevant images, and edit visuals quicker than ever before.
Trending searches
Beyond BIG Data
INTRODUCTION
IoT
1970s
The 70s
Reach: Specialists
Topology: Standalone Apps
Access: Centralized data-center
1980s
The 80s
Reach: Some Managers
Topology: App Suites and Integration
Access: The Local Area Network
1990s
The 90s
Reach: Most of the public
Topology: Websites and Web Portals
Access: WAN
2010
The 2000s
Reach: Things and People
Topology: Distributed Systems
Access: Mesh Networks
OLD
Features
Nothing goes away....
- Reports
- Dashboards
- Transactions
- Applications
- Web Sites
- Mobile Apps
Building on the OLD
NEW
- Now we have the ability to sense and respond to changes in the world
- Meet needs as soon as ... or even before they are noticed by humans
- We have reactive and proactive
City Snow Removal
Example
- Charging unit shows subtle pattern vibrations
- IoT system predicts battery charging failure in a couple of hours
- It messages driver to reroute truck to maintenance depot
- It messages another driver to pickup scheduled payload
- It automatically orders parts and sends them to maintenance depot
- Truck is put out of service
- Schedule adjusted for other trucks to deal with missing truck
BIG
DATA
DATA
GROWTH
Megabytes
MAINFRAME
GROWTH
Gigabytes
CLENTT/SERVER
Terabytes
INTERNET
Petabytes
CLOUD
Exabytes
IoT
SIZE
So how large are these data sizes?Gigabyte?Petabyte? Exabyte?
12
Terabyte 10
1 meter
15
Petabyte 10
1 km
18
Exabyte 10
1000 km
EXAMPLE
SIZE
- Average of half kilobyte per device per minute
- 50 billion devices by 2020
- half million minutes ter year
Thats 12.5 Exabytes per year
12,500 km
THE
LAKE
THE
LAKE
LIGHT
That's alot of data.
- tweets
- Social Media updates
- Amazon transactions
- News
- Marketing
This in/out processing works great for moderate IoT.... petabyte data sets
HEAVY
This is video content and works slightly different
- Original in the core and copies sent to the edge
- Enabled by content delivery networks
- Video close to where it is consumed
With IoT data will be kept at the edge and some of it sent back to the core
THE EDGE
HEAVY
These are things like
- Arduino Boards
- Raspberry PI Boards
- Edge Gateways
- Remote Servers
The furthest from the core where you can install a computer; you won't see these devices
AUTOMATION
AUTOMATION
REACH
Device out number people and their data is beyond human scale.
- Billions of sensors
- Millions of Edge Gateways
- Terabytes generated every day
- Exabytes datasets
Manual techniques won't work
SO MUCH DATA
How will we make use of it? How will we manage it?
- Automation
- AI or machine learning - extracting signals from noise
- Conversational Interfaces
AUTOMATION
THINGS
There are millions of these devices. We must change the way software is developed.
* Developers are responsible for the code they develop even when it is in operation
* they build models that drive fully automated operations
* ZERO touch is the main goal - complete automation
It is the only way to manage systems of this scale
INTELLIGENCE
ARTIFICIAL
Now we need to figure out what all this data means.This is where machine learning comes in.
It is a way of evolving algorithms to extract patterns.
It is important that we have human judgement to the machine learning else we will have to deal with ethical issues.
INTERFACES
CONVERSATION
These are conversational interfaces... bots that engage in conversations with users.
* Important in predictive systems
* inform you of predictions or events and then take action on your commands
* Smarter bots can explain WHY they came to a specific conclusion
These are the technologies that we need to bring to bear in the new paradigm of IoT data.