You are on page 1of 60

SEN1622 – Week 2.5 – 11.12.

2018

Technical Architectures
for IoT Services

Dr. Aaron Ding


Assistant Professor

1
Learning Objectives
• Describe enabling technologies across layers
• Identify trade-off and demands of IoT services
• Illustrate key features for edge computing
• Make trade-off between cloud and edge architectures
in designing IoT services

2
Outline

• Enabling Technologies
• Edge Architecture
• Takeaway

3
Outline

• Enabling Technologies
• Edge Architecture
• Takeaway

4
What’s Enabling IoT ?

5
Price

6
What About Technology ?

7
Computational Speed

8
Definitions
• Technology
– The application of scientific knowledge for practical purposes
– Devices developed from scientific knowledge

• Enabling technologies
– Can enable novel service offering
– Create organizational dependencies and financial costs

9
Why We Care About
Enabling Technologies ?

10
Example of Reading Digitalization
• Traditional books to digital booklet
– Easy to carry, update and index
– Small-size, inexpensive hardware
– New ecosystem for business

11
Example of Reading Digitalization

Computer industry
publishing and
Electronics Online
retailing

New ecosystem
collaboration
brings together
unrelated industries 12
Why Enabling Technologies?
• For your valorization plan
– How your service design includes enabling technologies?
– What generic elements to reuse (e.g., billing, authentication,
distribution)?
– Which stakeholders provide those generic elements?

• What you build can be considered as well


– Enabling technology for IoT ecosystem

13
How to Characterize ?

14
Layered Modular Architecture
• Hybrid of digital and physical design
– Interface between components
– Components are modular and independent
– Fusion of digital technology and physical products

15
Key Layers
• Device layer
– Physical capacity and computing

• Network layer
– Protocols and communication

• Application layer
– Content ownership and organization

• Platform layer
– Cross platform and distribution channel

16
Device Layer
• Computing devices
– Smart sensors, RFID tags, wearables, smartphones
– CPU, memory, storage, network interfaces

• Operating systems
– System software for hardware support
– Control and maintain physical components

17
Trade-offs for Device Layer
• What type of device to run the service?
– Smartphones
– Wearables
– AR/VR gadgets, HoloLens

• Does the service need to be mobile or static?

• Is there physical restriction?


– Target scenario: home, office, industry factory, air, underwater
– Power supply: availability, charging
– Network connectivity: fixed, wireless, range, none

18
Network Layer
• Physical transmission
– Wired: ethernet cables, fibers
– Wireless: Cellular 3G/4G/5G, WiFi, Bluetooth

• Logical transmission
– Protocol standards: TCP, IPv6

19
Trade-offs for Network Layer
• What network performance is needed?
– Throughput
– Latency
– Range
– Mobility support
– Transmission power

• Is network connectivity always required?


– What if outside coverage
– Roam charges abroad

20
Application Layer
• Content display and ownership
– Interaction logic with users: input and output
– Data access and storage: local or remote

• Service function organization


– Thin client, depending on cloud or other clients
– Heavy client, stand-alone
– Client-Server paradigm
– Peer to Peer (P2P)

21
Trade-offs for Application Layer
• How to interact with users?
– Multi-touch display, physical button
– Gesture recognition, voice

• Where to store data?


– Local
– Cloud

• How to organize the service logic?


– Centralized
– Decentralized

22
Platform Layer
• Platform ecosystem
– Development toolkit, community, popularity
– Deployment cost

• Distribution channel
– Application store, central repository
– Promotion strategy

• Service lock-in
– Platform dependent or cross platform
– Function virtualization

23
Trade-offs for Platform Layer
• Which platform ecosystem?
– Device support
– Application diversity, focus

• Is App Store needed?


– Cost
– Visibility

• Is cross-platform required?
– Native vs. virtualized

24
What About IoT Demands ?

25
Demands of IoT
• Hardware dependent
– Size factor
– Price/Performance

• Domain specific
– Infrastructure
– Regulations

• Service oriented
– User experience
– Performance

26
Demands of IoT
• Key requirements for IoT enabling technologies
– Cost efficiency
– Energy efficiency
– Latency
– Resilience
– Extensibility
– Scalability
– Multi-tenancy
– Security and Privacy

27
Summary: IoT Enabling Technologies

Layer Technologies Demands


Device Layer Wearables, Drones, Cost Efficiency
AR/VR, Connected Cars Energy Efficiency

Security & Privacy


Network Layer 5G, BLE, WiFi, Zigbee, Latency
VLC, TLS/SSL, IPv6 Resilience

Application Layer HCI, Voice control, P2P Extensibility

Platform Layer Virtualization, Service Scalability


distribution stores Multi-tenancy

Which stakeholder will use which element? Why and How?


28
Outline

• Enabling Technologies
• Edge Architecture
• Takeaway

29
Architecture for IoT

30
Cloud for IoT
• Traditional partner
– All for one, one for all
– Power of centralization

• Mature business model


– Data driven
– Comprehensive ecosystem
– Popular and dominant

31
Cloud Advantages
• Computing paradigm

32
What About Data Privacy ?

33
What We Have Now

34
When IoT Really Comes
• Internet of Too Many Things

35
Edge Architecture

36
5G Trend
• Pervasive
– Coverage
– Service diversity
– Mobility support

• Performance
– 20 Gbps
– 1ms latency

37
Empowering IoT

A. Y. Ding, M. Janssen. Opportunities for Applications Using 5G Networks:


Requirements, Challenges, and Outlook. In Proceedings of the 7th International
38
Conference on Telecommunications and Remote Sensing (ICTRS '18).
Why the Edge?
• Data pattern changed

• New demands

• Dilemma of the cloud


Edge Ecosystem

Deployment of IoT
Cyber-Physical
Systems

39
Car Data

40
Safety in Driving
• Tesla autopilot/self-driving accident on 7th May 2016

41
Use Case

ices
S erv
r o
e Mic
E dg

42
Smart Transport
• Lamppost 4.0
– Road safety enhancement
– Edge data processing

• Benefits
– Latency
– Energy
– Congestion reduction
– Privacy

43
Edge Architecture for IoT

Edge Analytics and Services

Edge Edge Edge


Computing Communication Security

Architecture and Platform

Cyber-Physical Infrastructure

44
Scenarios

45
IoT Really Needs Edge?
• Existing challenge
– IoT and Cloud are already in use
– 5G is coming !

46
IoT Really Needs Edge?
• Existing challenge
– IoT and Cloud are already in use
– 5G is coming !

• New challenges
– Data growth
– Privacy concern
– Delay sensitive applications: AR/VR, car services
– Uplink backhaul: 5G is not enough

47
Key Features for Edge
• Locality
– Data source, end users
– Congestion avoidance

• Latency

• Resilience

• Privacy Enforcement

48
But How ?

49
Edge Computing in Practice
• Fine-grained offloading for smart infrastructure

Lightweight Virtualization

50
Lightweight Virtualization for Edge

51
Mobile Oriented Design
• Cloud based design

52
Dedicated for IoT
• Reverse direction
IoT

53
Edge Offloading
• Cloud – Edge – IoT
Edge IoT
Offloading

54
The Real Benefits

55
Edge or Cloud ?

56
Cloud vs. Edge
Features Cloud Edge
Data Management Centralized Decentralized

Computing Capacity High, good for heavy load Low, but growing fast

Storage Capacity Sufficient Limited

Network Latency High latency Low latency

Resilience Bottleneck, central point Distributed, robust


of failure
Multi-tenancy Yes and scalable Partially

Privacy Control Depending on service Good level of control


contract

57
Outline

• Enabling Technologies
• Edge Architecture
• Takeaway

58
IoT Needs Both Edge and Cloud
• Enabling technologies
– Layers of device, network, application, platform
– Trade-off across layers

• Architectures for IoT services


– Motivation and features of edge
– Trade-off between edge and cloud for IoT service design

59
Reference
• “Information technology - Internet of Things - Definition and Vocabulary”, ISO/IEC JTC 1
N13023
• Managing IoT at the Edge: The Case for BLE Beacons. SmartObjects @ MobiCom ’17
• Demo: iConfig: What I See is What I Configure. CHANTS @ MobiCom ’17
• M. Haus, A. Y. Ding, C. Xu, J. Ott. 2018. Touchless Wireless Authentication via LocalVLC.
In Proceedings of the 16th Annual International Conference on Mobile Systems,
Applications, and Services (MobiSys '18).
• M. Yannuzzi et. al. A New Era for Cities with Fog Computing. IEEE Internet Computing,
2017
• Roberto Morabito, Vittorio Cozzolino, Aaron Yi Ding, Nicklas Beijar, Joerg Ott.
Consolidate IoT Edge Computing with Lightweight Virtualization. IEEE Network, Vol. 32,
No. 1, pp. 102-111, 2018.

60

You might also like