IoT Network Research at LIG Drakkar Team Andrzej Duda LIG - - PowerPoint PPT Presentation

Andrzej Duda LIG Grenoble-INP - Ensimag

IoT Network Research at LIG


Drakkar Team

Overview

• Drakkar Team
• IoT - comparison of energy

consumption in IoT networks

• Issues in capillary IoT

networks - routing

• Issues in cellular IoT networks
• access methods
• Security
• Reproducibility - WalT platform

2

Example TI MCU

5 10 15 20 TX/RX MCU Sensors Sleep

Current (mA) ~µA

Drakkar Team

3

Drakkar Team

• Head
• Andrzej Duda, PR Grenoble INP
• Permanent staff
• Olivier Alphand, MC Grenoble INP
• Etienne Dublé, IR CNRS
• Martin Heusse, PR Grenoble INP
• Franck Rousseau, MC Grenoble INP
• Pascal Sicard, MC UGA
• Bernard Tourancheau, PR UGA
• 14 PhD students

4

"That is about as fun as herding cats"

Main domains

• Wireless networks
• WLANs 802.11, advanced MAC
• IoT - Sensor & actuator networks
• all-IP

, LPWAN

• MAC, routing, data-centric
• Security, traffic analysis
• anomaly detection, DNS cybersecurity
• security of IoT

5

Data rate Communication range 10 m 100 m 1 km 10 km WSN ZigBee/15.4 Bluetooth WLAN Wi-Fi/ IEEE 802.11 kb/s Mb/s Cellular 3G/UMTS 4G/LTE Long Range SigFox/Lora 5G-MTC

Main results

• 802.11 Performance Anomaly
• INFOCOM 2003
• Idle Sense, an optimal 802.11 access method
• ACM SIGCOMM 2005
• Recent INFOCOM papers
• Wake-on-Idle
• Detecting applications in encrypted flows
• IEEE TMC, ACM CCR
• TPC
• Infocom

6

IoT - comparison of energy consumption in IoT networks

7

Capillary vs. Cellular

8

Internet Gateway

IP mesh

Short range capillary network Internet Base Station

Proprietary Protocols e.g. LoRa, SIGFOX

Long range one-hop network

Industrial IoT, e.g. 802.15.4 TSCH

Energy consumption in 802.15.4

9 Beacon

Inactive Period

CAP

DATA S I F S ACK

SuperFrame Duration (ton) Beacon Interval (t )

I F S S/L

DATA

DATA S I F S ACK I F S S/L Beacon Backoff (0) CCA DATA I F S S/L Tx Rx Idle Sleep Backoff (0) CCA Backoff (0) CCA

CI

• Energy consumption

E(t) = X

S

PS × tS, S ∈ {T x, Rx, Idle, Sleep}

Example TI MCU

5 10 15 20 TX/RX MCU Sensors Sleep

Current (mA) ~µA

Lifetime, 1 pkt per day

10

5 10 15 20 25 30 35 10 100 1000 10000 0.000926 0.009259 0.092593 0.925926 Lifetime in years sa (bytes) Application Throughput ra (bps)

BLE 4.2 (1 Mb/s) 802.15.4 (250 kb/s) TSCH (250 kb/s) 6-State Model TSCH (250 kb/s) 802.11 PSM (11 Mb/s) ah: MCS10,1Mhz (150 kb/s) ah: MCS8,2Mhz (7.8 Mb/s) ah: MCS9,16Mhz (78 Mb/s) LoRa EU Min (250 b/s) LoRa EU Max (11 kb/s) SIGFOX EU Min (100 b/s) SIGFOX EU Max (1000 b/s) 1% DC curves are dashed, 0.1% DC are full

BLE/15.4 LoRA SIGFOX 802.11ah

• Snow level metering (50B/day)

Lifetime, 1 pkt per 1s

11

BLE 4.2 (1 Mb/s) 802.15.4 (250 kb/s) TSCH (250 kb/s) 6-State Model TSCH (250 kb/s) 802.11 PSM (11 Mb/s) ah: MCS10,1Mhz (150 kb/s) ah: MCS8,2Mhz (7.8 Mb/s) ah: MCS9,16Mhz (78 Mb/s) LoRa EU Min (250 b/s) LoRa EU Max (11 kb/s) SIGFOX EU Min (100 b/s) SIGFOX EU Max (1000 b/s) 1% DC curves are dashed, 0.1% DC are full

0.01 0.1 1 10 100 10 100 1000 10000 0.08 0.80 8 80 Lifetime in years (logscale) sa (bytes) Application Throughput ra (kbps) BLE 802.11 PSM 15.4e TSCH 802.11ah

• Heart rate monitor (50B/s)

Issues in cellular IoT networks - routing

12

Issues in Capillary

• Needs a routing layer for IP mesh
• RPL (Routing Protocol for Low power and Lossy Networks)
• IETF standard
• LRP (Lightweight Routing Protocol) - enhanced RPL

13

Internet Gateway

IP mesh

Short range capillary network

LRP in a word


Local Repair

14

RPL


Protocol Background Traffic

15

RPL — 1 sink, 40 nodes

LRP


Protocol Background Traffic

16

LRP — 1 sink, 40 nodes

Issues in cellular IoT networks - access methods

17

Issues in Cellular

18

Internet Base Station

Proprietary Protocol e.g. LoRa

Long range one-hop network

• Needs a scalable access method
• LoRa, SIGFOX - no access method (ALOHA), limited traffic (e.g.

1 pkt. per day)

• 5G MTC - what access method?
• need for scalable access

NS-3 module for LoRa

19

• Take into account capture effect - lower loss rate than in ALOHA

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 100 200 300 400 500 600 700 800 900 1000 Packet loss/collision ratio Number of Nodes

Simulation, packet loss Simulation, collision ratio Measured packet loss [8] Measured collision ratio [8]

NS-3 module for LoRa

20

• CSMA results in better packet delivery ratio
• Takes away duty cycle restrictions of ISM 868 band

0.0 0.2 0.4 0.6 0.8 1.0

Packet Delivery Ratio Number of Nodes

LoRaWAN CSMA CSMA-10

5G MTC

• Need to find a suitable MAC
• fits 5G architecture
• light signaling
• low energy
• massively scalable

21

IoT Security

22

IoT Object Security

23

IoT Object Security

24

Authorization Blockchain

Proxy server Resource Servers

Request keys R e q u e s t e x e c s m a r t c

• n

t r a c t Verify if client’s token exist in blockchain Fetch data

Clients

2 4 5 7

Receive keys

6

Token generation for client

1 Resource Owner publishes

smart contract

3

Resource Owner Key Servers

Reproducibility - WalT platform

25

Reproducibility

• Validating new protocols for wireless networks is a

challenging task

• simulations far from realistic conditions
• Perform real-world experiments!
• Reproducibility - when an experiment can be reproduced

under different conditions, while providing sufficiently similar results

• reproduce experiments, build upon, and compare their

results with the previous work

26

"I could not imagine much progress by reading only, without experimental facts and trials", M. Faraday

WalT - reproducible platform to run reproducible experiments

27

• http://walt.forge.imag.fr

Use WalT to monitor Sensor Network

28

• 20 nodes deployed in LIG
• Used for various demos (in a backpack)

Conclusions

• Cellular vs. Capillary IoT
• several technologies available optimized for specific use

cases

• 5G MTC
• a lot of current research
• massively scalable MAC
• low-latency MAC
• Importance of experimentation
• reproducibility
• http://walt.forge.imag.fr

29

A conclusion is the place where you got tired of thinking.