Architecture of Industrial IoT - ASRENasrenorg.net/eage2016/sites/default/files/files/Marc...

Architecture of Industrial IoTDecember 2, 2016

Marc Nader@mourcous

Copyright 2015 DATACONSULT All rights reserved. No part of this presentation in all its property may be used or reproduced in any form without a written permission 2

Branches of IoT

IoT

Consumer IoT(Wearables, Cars, Smart

homes, etc.)

Industrial IoT (IIoT)

Smart Gateways

Wireless Sensor

Networks

Mesh(Zigbee,

Dust, Wifi)

LPWAN (LoRa,

NB-IOT, Sigfox)


The IOT Paradigm

How do we get there?

HOW?

http://mms.businesswire.com/media/20150929006449/en/488323/5/Nexus_6P-Frost-Front_and_Back.jpghttp://mms.businesswire.com/media/20150929006449/en/488323/5/Nexus_6P-Frost-Front_and_Back.jpg


IOT Data Flow

Capture CodeTransmitRAN

TransportIP, MPLS, etc.

Store Analyze Action

Site Network Data Center/Cloud

Big Data /Analytics /

Application builders

IoT Enabler

Copyright 2016 DATACONSULT All rights reserved. No part of this presentation in all its property may be used or reproduced in any form without a written permission

IOT Data Flow

Within the Site

5

Part of Operation Technology (OT);Sensors with analog (0-10V) or digital electrical outputs;Powered by the IOT node or through an external source;The IOT node transforms the electrical signal into data packets;Message Queuing Telemetry Transport (MQTT),

Constrained Application Protocol (CoAP)Rugged design;Or Sensors can take form of any IP based protocol.

Capture Code

IoT nodes connect one or multiple sensors and translate an analog signal into packets.

Sensors IoT nodes

TransmitRAN

TransportIP, MPLS, etc.



Low Power Wide Area (LPWAN)

Po

we

r u

sag

e

Range >10km


IOT Data Flow

Radio Access Network: Transmitting the M2M chatter

7

Collecting data from on a massive scale while preserving the sensor battery life is a challenge;Ferocious competition for the Low Power Wide Area (LPWA) technology dominance;

The "LPWAR": lets look at the market alternatives in this area:Sigfox: Global IoT Operator.LoRa: Enable operators and enterprises to create IoT networks3GPP: NB-IOT, LTE-M, NB-LTE & 5G: Natural evolution of mobile operators.

Fixed & Short rangeWifi + 3/4G/FixedZigbee + 3/4G/Fixed

Capture Code Transmit Transport

Sensors IoT nodes IoT Gateways IP Network



Sigfox

Privately owned, French.

Ultra narrow band 868Mhz - 100Hz (BPSK)140 messages per unit per day. 1 every 10 minutes.

12 bytes per message15 km, bi-directional.High battery life (Up to 20 years)

Limited bandwidth applicationsCity wide deployments, short message

communicationsShipping and very activeNetwork as a service model.Cloud based.

Technology Applications


LoRa

LoRa Alliance, backed by Cisco, IBM,...

Star of Stars topology, Encrypted protocolWideband CDMA: 868Mhz -


Narrow Band IOT - NB - IOT

3GPP, Huawei (Neul acquisition)

LTE-M - 1.4Mhz (1Mbps/1Mbps)NB LTE-M - 200khz (200kbps/170kbps)One LTE cell can handle 150k LTE-M devices10 years battery life (200 bytes daily update)Low cost terminal

Leverage the LTE network with a software upgrade.LTE-M +15dB link budget (x7 coverage)NB LTE-M +20dB link budget (x10 coverage)

Licensed spectrum3GPP Rel. 13 NB-IOT Coming with LTE-A Pro (4.5G)Mobile operators become IoT operators with

a software upgrade..



Fixed & Short Range IoT Connections


Fixed or Short Range Wireless


WiFi / Zigbee / Dust + 3G/4G

Local Aggregation of IOT nodes is done through:WifiZigbee / 802.15.4 / Dust / Z-Wave

Site aggregation is done through 3G/4G or fixedNeed external power source for the aggregationHigher throughput

High throughput, sending voice/videoReal Time or near-RTIn home, in plant, on campus.



IOT Data Flow

Transmitting the M2M chatter

14

Capture Code Transmit Transport

Technology decision factor




Evolutions of Apps

Making Sense of Data


IoT Applications

The three layered approach


IOT Data Flow

Gathering, structuring and processing the data

17

Input from IOT Gateways (MQTT)Input from IOT Network Server (kafka)Any other inputData modeling and data set definitionsMachine Learning, Analytics

Data Science for IOT: making sense ofmassive data and give the machines theability to take intelligent decisions.

Capture Code Transmit Transport Store Analyze Action



The Challenge of Data in Motion

Data at Rest Data in MotionVS

Capture Analyse Actiont1 t2

Capture

Analyse

Action

rt, nrt


Hardware/Software Platform Summary

Long Range

Short Range

IoT Gateway

Node & Sensor

Node + Sensors

Cellular

Fixed IoT Network Platform

Application #1

Application #2

Application #3


Hardware/Software Platform Summary

Long Range

Cisco LoRa Gateway

Node & Sensor

Node + Sensors

Ogero Backbone

Actility Network Platform

Application #1

Application #2

Application #3Short Range

MIC, AS128 Encryption (NwkSkey, AppSkey) HTTPS Post, MQTT, Kakka

AS256 IPSec - Cisco

Ogero Network

Optional Payload Encryption (App developers)


Full Network Design


The IoT Services Framework

Connectivity Management

Device Management

Data Management

S

E

C

U

R

I

T

Y

Application

Enterprise

Integration

Cloud 2 Cloud


Fog Computing

Eliminating the unnecessary chatter on the radio

23


Fog Computing

Cloud Computing

Preprocessing of data from sensors saves on transmissions costs.

Bringing some intelligence closer to the edge.

Computing on the IOT gateway

(managing different nodes)

Decreased network chatter and cloud storage. Linux IoT nodes, complex computing done local.


Mist Computing

Eliminating the unnecessary chatter on the radio

24


FogCloud Computing

Bringing some intelligence closer to the edge.

MistComputing

Computing on the IOT node

(sensor or actuator)

Decreased latency More autonomy


A Step In the Right Direction

Interoperability efforts


Remember Stuxnet


The Security Paradigm

IOT Gateways(Linux, cheap chips)

How about security?

https://upload.wikimedia.org/wikipedia/commons/thumb/f/fc/Project-triangle.svg/2000px-Project-triangle.svg.pnghttps://upload.wikimedia.org/wikipedia/commons/thumb/f/fc/Project-triangle.svg/2000px-Project-triangle.svg.png


The End of The Fortress Approach

Once the hacker is in, Game Over!

Intelligence is distributed and networks are intertwined.

The Goal is not to keep out only.Detect, Isolate, Heal


You are as secure as your weakest link


The Zero-Trust Network

30

Untrusted TrustedUntrusted

https://pixabay.com/static/uploads/photo/2016/03/21/23/43/computer-1271863_960_720.pnghttps://pixabay.com/static/uploads/photo/2016/03/21/23/43/computer-1271863_960_720.pnghttps://pixabay.com/static/uploads/photo/2016/03/21/23/43/computer-1271863_960_720.pnghttps://pixabay.com/static/uploads/photo/2016/03/21/23/43/computer-1271863_960_720.png


Proposed IOT Security Framework

Per device:Authentication: endpoints should be fingerprinted.Authorization: establishing the cross platform trust

relationships.Network Enforced Policy: all elements that route and

transport endpoint traffic securely over the infrastructure.

Analytics: Data, generated by the IoT devices, is only valuable if the right analytics algorithms or other security intelligence processes are defined to identify the threat.

Cisco Securing the Internet of Things: A proposed Frameworks.Google BeyondCorp


IOT

Finally

IOTInformation Technology

Operation Technology

Business Outcome

Operation technology & Low Current Radio Access Networking Data Integration, software Data science

System Integration

http://ian.umces.edu/imagelibrary/albums/userpics/101505/normal_ian-symbol-dollar-sign.pnghttp://ian.umces.edu/imagelibrary/albums/userpics/101505/normal_ian-symbol-dollar-sign.png

Thank you!

Marc [email protected]

@mourcous
mailto:[email protected]

Architecture of Industrial IoT - ASRENasrenorg.net/eage2016/sites/default/files/files/Marc...

Documents

Transcript of Architecture of Industrial IoT - ASRENasrenorg.net/eage2016/sites/default/files/files/Marc...