ds2os.org/

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Securing the Internet of ThingsMarc-Oliver Pahl (TUM) | s2labs.org | pahl@tum.de

2Slabslabslabs

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Why do we need a secure IIoT?

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemUser (un)aware monitoring

3Photo courtesy Universal Robots A/S

What?How often?

Weight?

Preferences?

Who?

Where?

With whom?

When?

@work

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Foto:Steve Jurvetson (Flickr) https://www.flickr.com/photos/jurvetson/7408451314

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemSecurity Challenges include

PrivacyReliability

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

How can we achieve a secure IIoT?

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemThe usual suspects

• Authentication • Authorization • Confidentiality

Entities

Devices

Services

Users

LinksData

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Why is it challenging?

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Heterogeneity

Imag

e So

urce

: Pra

shan

t Sha

rma,

Mic

roso

ft Ve

ntur

es |

http

s://w

ww.

mic

roso

ftven

ture

s.co

m/b

log/

entry

/Inte

rnet

ofTh

ings

101U

nder

stan

ding

theB

uild

ingB

lock

softh

eCon

nect

edW

orld

%7C

1064

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

10

ComplexityFlickr | Mark Skipper | https://www.flickr.com/photos/bitterjug/7670055210

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemWhom to trust?

• Hardware?

• Software Services?

• My answer: no-one but the components you design as enabler for the IoT: MIDDLEWARE.

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Security by-design

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Data Centric IoT

Smart SpaceKnowledge Agent

Context Repository

Context Manager

Adaptation

ActuatorSensor

Adaptation Adaptation

Service UI Service Service Service

Heterogeneous Smart Devices

Bidirectional Adaptation

Context Management Virtual State Layer

Orchestration Workflows, etc.

PeopleInterface Devices

Interface Devices

Physical World

Phys

ical

Wor

ld

Dev

ices

Serv

ices

Con

text

Data Centric IoT

VSL Knowledge Agent

Knowledge Repository

Knowledge Manager

Adaptation

ActuatorSensor

Adaptation Adaptation

subscribe notifygetset

Service UI Service Service Service

subscribe notifygetset

virtual nodes

Heterogeneous Smart Devices

Bidirectional Adaptation

Context Management Virtual State Layer

Orchestration Workflows, etc.

PeopleInterface Devices

Interface Devices

Physical World

Leve

l of A

bstra

ctio

n

Phys

ical

Wor

ld

Dev

ices

Serv

ices

Con

text

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Data

Entity Management

Managed Entities

Comprehensive Management

User Interfaces

A Data Centric IoT

Needs autonomous management!

The Virtual State Layer

Services

VSL Overlay

Hardware Underlay

Other Services

Gateway Services

Logical Connectivity

Physical Connectivity

Knowledge Agent

Gat

eway

Adv

ance

d R

easo

ning

Orc

hest

ratio

n

Use

r In

terf

ace

...

ActuatorSensor

DS2OS Site Local Global

Cen

tral

Mod

elR

epos

itory

App

Sto

re

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Smart IoT Devices1

Man

aged

IoT

Spac

e

(Adaptation) Services2a

(Orchestration) Services2b

VSL: Virtual Objects3

Glo

bal R

epos

itory

Central Model RepositoryData Models

4

host5

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

How to achieve Security-by-Design?

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemTwo approaches

1. Handling security in the middle in a non-circumventable way

2. Retrofitting Security

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemFor me everything is a service

Development Distribution

Configuration

Deployment

Update

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem“Crowdsourced” Development

Edge

Internet

IoT space

IoT space

IoT space

IoT space

IoT spaceIoT

space

StoreD

D

DevelopersUsers

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

1. Handling security in the middle in a non-circumventable way

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Distributed Smart Space Orchestration System

µService Store

IoT Smart Space

Global

KA KA

µS

SLSM

config

µSµS µS

µService Package

executable

manifest

Context Model Repo

e

c

1

3

4

a

2

NLS

Mb

VSL Middleware

fd

KA

KA

*LSM {Site,Node}-Local Service Manager

Computing Node

Middleware Interface

Context Model (Data Model)

u

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Distributed Smart Space Orchestration System

service package

metadata

store

IoT Site

SLCA

global

cert

KA KA

svc

SLSM svcsvc svc

executable

cert

e

cs

1

2

3

a

NLS

M

b

VSL Middleware

fd

KA

KA

SLCA Site-Local Certificate Authority

cert

*LSM {Site,Node}-Local Service Manager

Computing Node

Middleware Interface

Service Certificate

Private Key

Signature

u CA

cert certcert

cert cert cert

Z

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemDistributed Revocation via Short Lifetime

Certificates and fully automated Renewal

0

50

100

150

200

250

300

350

400

450

500

550

0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000

AverageTraffc

(Bytes/s)

Certifcate Lifetime (seconds)

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemCosts: Energy

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

00:00 03:00 06:00 09:00 12:00 15:00 18:00 21:00 24:00 27:00 30:000

5

10

15

20

25

30

35

40InputC

urrent(A)

CPU

Usage

(%)

Time

CPU usageInput Current

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

2. Retrofitting Security

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemApproach in a

• Blackbox assumption

• Passive traffic monitoring

• Behavior modeling

• Anomaly detection

• Firewalling

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemApproach

Raspberry Pi 3 Model B V1.2

Power

HDMIAudio

USB

2x

USB

2x

ETH

ERN

ET

DSI

(DIS

PLAY

)

CSI (CAM

ERA)

GPIO

© Raspberry Pi 2015

µS

Raspberry Pi 3 Model B V1.2

Power

HDMIAudio

USB 2x

USB 2x

ETHERN

ET

DSI (D

ISPLAY)

CSI (

CAM

ERA

)

GPIO

© Raspberry Pi 2015

µSµS

µS

µS

µSService Communication Monitor, Analyzer, and Firewall

MicroserviceInter-Node Comm. Interface

Service Runtime EnvironmentµS Model Federation Service

µService Store

Raspberry Pi 3 Model B V1.2

Power

HDMIAudio

USB 2x

USB 2x

ETHERN

ET

DSI (D

ISPLAY)

CSI (

CAM

ERA

)

GPIO

© Raspberry Pi 2015

µS

IoT Site

VSL Middleware

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemWho talks to whom?

0 2 4 6 8 10 12 14Time (m)

0

5

10

15

Num

ber o

f edg

es a

nd v

ertic

es

First learning phaseLearning phase at the addition of servicesAnomalous behavior needing userNumber of verticesNumber of edges

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemClustering Periodicities

-20 -15 -10 -5 0 5 time (s)

pack

et a

rriva

l

Outgoing traffic from a washing machine service

Write to battery1Read to battery1Write to battery2Read to battery2Read to thermometerRead to movement

-20 -10 0 time (s)

0

0.5

1

1.5

2

2.5

inte

r-arri

val d

urat

ion

(s)

Battery read

-20 -10 0 time (s)

0

0.5

1

1.5

2

2.5

inte

r-arri

val d

urat

ion

(s)

Battery write

-20 -10 0 time (s)

0

0.5

1

1.5

2

2.5

inte

r-arri

val d

urat

ion

(s)

Thermometer write

-20 -10 0 time (s)

0

2

4

6

inte

r-arri

val d

urat

ion

(s)

Movement write

-20 -10 0 time (s)

0

0.5

1

1.5

2

2.5

inte

r-arri

val d

urat

ion

(s)

All the traffic

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stemQuality

Classification errors over time

DoS attack

service update

DoS attack

anomaly value too low

0 5 10 15 20Time in hours

0

50

100

150

200

250

300

350

Num

ber o

f cla

ssifi

catio

n er

rors

0

69800

139600

209400

279200

349000

Tota

l num

ber o

f pac

kets

False PositiveFalse NegativeTotal number of packets

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Summary

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Distributed Smart Space Orchestration System

µService Store

IoT Smart Space

Global

KA KA

µS

SLSM

config

µSµS µS

µService Package

executable

manifest

Context Model Repo

e

c

1

3

4

a

2

NLS

Mb

VSL Middleware

fd

KA

KA

*LSM {Site,Node}-Local Service Manager

Computing Node

Middleware Interface

Context Model (Data Model)

u

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

[1] M.-O. Pahl and F.-X. Aubet, “All Eyes on You: Distributed Multi-Dimensional IoT Microservice Anomaly Detection,” presented at the 2018 14th International Conference on Network and Service Management (CNSM) (CNSM 2018), Rome, Italy, 2018.

[2] M.-O. Pahl and L. Donini, “Securing IoT Microservices with Certificates,” presented at the Network Operations and Management Symposium (NOMS), 2018.

[3] M.-O. Pahl and M. Loipfinger, “Machine Learning as a Reusable Microservice,” presented at the Network Operations and Management Symposium (NOMS), 2018.

[4] M.-O. Pahl, F.-X. Aubet, and S. Liebald, “Graph-Based IoT Microservice Security,” presented at the Network Operations and Management Symposium (NOMS), 2018.

[5] F.-X. Aubet, M.-O. Pahl, S. Liebald, and M. R. Norouzian, “Graph-based Anomaly Detection for IoT Microservices,” presented at the Passive and Active Measurement Conference (PAM), 2018.

[6] M.-O. Pahl, G. Carle, and G. Klinker, “Distributed Smart Space Orchestration,” presented at the Network Operations and Management Symposium 2016 (NOMS 2016) - Dissertation Digest, 2016.

[7] M.-O. Pahl, “Data-Centric Service-Oriented Management of Things,” presented at the Integrated Network Management (IM), 2015 IFIP/IEEE International Symposium on, Ottawa, Canada, 2015, pp. 484–490.

[8] M.-O. Pahl and G. Carle, “Crowdsourced Context-Modeling as Key to Future Smart Spaces,” presented at the Network Operations and Management Symposium 2014 (NOMS 2014), 2014, pp. 1–8.

[9] M.-O. Pahl and G. Carle, “Taking Smart Space Users into the Development Loop: An Architecture for Community Based Software Development for Smart Spaces,” presented at the Proceedings of the 2013 ACM Conference on Pervasive and Ubiquitous Computing Adjunct Publication, New York, NY, USA, 2013, pp. 793–800.

[10] M.-O. Pahl and G. Carle, “The Missing Layer - Virtualizing Smart Spaces,” presented at the 10th IEEE International Workshop on Managing Ubiquitous Communications and Services 2013 (MUCS 2013, PerCom 2013 adjunct), San Diego, USA, 2013, pp. 139–144.

[11] M.-O. Pahl, H. Niedermayer, H. Kinkelin, and G. Carle, “Enabling Sustainable Smart Neighborhoods,” presented at the 3rd IFIP Conference on Sustainable Internet and ICT for Sustainability 2013 (SustainIT 2013), Palermo, Italy, 2013.

[12] M.-O. Pahl, C. Niedermeier, M. Schuster, A. Müller, and G. Carle, “Knowledge-based middleware for future home networks,” presented at the WD'09: Proceedings of the 2nd IFIP conference on Wireless days, Paris, France, 2009.

ds2os.org/

Orc

hest

ratio

nD

istr

ibut

edSm

art

2pac

eSy

stem

Securing the Internet of ThingsMarc-Oliver Pahl (TUM) | s2labs.org | pahl@tum.de

2Slabslabslabs