2. • Practical IoT - What is it? Using the AWS
IoT as an example
• IoT in action - a Bath Car Parks demo
• Discussion
3. • What is AWS IoT...?
• Message Broker - MQTT and HTTP topic-
based pub/sub model
• Rules Engine - SQL-based language to
select data from message payloads and
push it into S3, DynamoDB and Lambda
• Thing Registry - Organises the resources
associated with each thing, including
certificates
4. • Thing Shadows service - Provides
persistent representations of your
things in the AWS cloud
• Thing Shadow - JSON document used
to store and retrieve current state
information for a thing (device, app,
and so on)
• Device Gateway—Enables devices to
securely and efficiently communicate
with AWS IoT
7. • Demo developed primarily to show
capability of the IoT platform...
• ...but nothing here that couldn't also
have been developed directly on the
underlying data
• Don't get too excited - AWS IoT is
designed to handle billions of devices -
this demo deals with 8 car parks
8. • Single EC2 instance (virtual server) in
AWS simulating 8 devices ('things')
• ...one for each car park in Bath
• Each 'device' pulls open data from the
BathHacked data store every 5
minutes providing information about
how full each car park is
• This data is then published to the AWS
IoT platform using secure MQTT
9. • The Rules Engine parses each message
and pushes a count of the number of
available spaces into a DynamoDB table
• A central control process monitors the
last 20 minutes of data from each car
park in DynamoDB and assesses
whether it is filling or emptying and how
long until it is full
• A short message is then pushed back to
the 'device' via the Message Broker
using MQTT
10. • The same message is optionally
published to Twitter (see
@bathcarparks)
• Finally, a website shows historical data
- http://www.bathcarparks.info
11. • Future work...
• Scope for added features
• Next nearest car park with spaces
• Typical spaces at given time/day
• Main enhancement would be to use AWS
IoT as primary collection mechanism
• Which opens up a debate about why use
IoT rather than existing methods?
12. • Conclusions...
• Well, the protocols work and integrate well
with the rich service environment offered by
AWS
• Coding devices is reasonably straight-forward
• Device SDK and language support somewhat
limited but evolving
• Arduino and Rasberry Pi obvious candidate
devices - various starter kits available
• Managing registry at scale will require
automation
14. • Discussion…
• Does IoT offer anything new?
• Is the cost/practicality of deployment at
scale too high?
• What challenges are there around getting
hold of IoT data sources?
• What are the issues around ownership and
IPR?
• What are the benefits over current
approaches?