Iot Network Architecture and Design: Mms Institut Teknologi Del

IoT Network Architecture and
Design
MMS
Institut Teknologi Del
Overview
• Why?
• Drivers Behind new Network Architecture
• Comparing IoT Architecture
• Simplified IoT Architecture
• Core IoT Functional Stack
• IoT Data Management and Compute Stack
Why?
• Everything need a plan
• IoT Network implementation differ from Traditional
IT which needs new approach in planning and
implementation
Drivers Behind New Network
Architecture
• OT (Operational Technology) networks, which is the target area of IoT
Implementation, have unique characteristics and constraints that are
not easily supported by traditional IT network architectures.
• In General, they key different between traditional IT and IoT is the data
• Traditional IT concerned with a reliable and continuous support of business
application (email, web, ERP)
• IoT is all about the data generated by sensors, how it is used, transported,
collected, analyzed and acted upon
• 5 main drivers:
• Scale
• Security
• Constrained device and networks
• Data
• Legacy Device Support
Scale
• Millions of endpoints in IoT network need different
approach in implementation and operation
compared to thousands of nodes in traditional IT
networks
• E.g. IPv4 address space exhaustion is unable to
meet IoT scalability requirement.
• Solution is to use IPv6
Security
• Case: at 3:30 p.m. on December 23, 2015, the Ukrainian power grid
(SCADA System) experienced an unprecedented cyber attack that
affected approximately 225,000 customers.
• IoT devices which often physically exposed introduce a new challenge
in designing security strategy.
• Requirements for optimum IoT system security:
• Authentication (Be able to identify and authenticate all entities involved in the
IoT service)
• Encryption (Ensure that all user data shared between the endpoint device and
back-end applications is encrypted)
• Secure data storage (Comply with local data protection legislation so that all
data is protected and stored correctly)
• Utilize an IoT connectivity management platform and establish rules- based
security policies.
• Take a holistic, network-level approach to security (Firewall, WAF, IDS/IPS)
Constrained Devices and Networks
• Most IoT sensors are designed for a single

job, and they are typically small and
inexpensive
• Limited power,
• Limited CPU
• Limited Memory
• Limited data transfer capability.
• IoT requires a new breed of connectivity ESP32 SoC (system on chip):
- 80 to 240MHz CPU,
technologies that meet both the scale and - 4MB Flash,
constraint limitations - 520KB RAM,
- low power sleep,
- ADC, SPI, I2C, UART
Data
• Data is generated by millions of nodes

• How to transport data from thousands – millions of
nodes to data center?
• Unstructured format, how to to store and analyze the
data?
Legacy Device Support
• Different from traditional IT Network devices, IoT

end devices are likely to be on the network for a
very long time (decades)
• In the era of bluetooth, wireless sensors, IoT
infrastructure often have to support device based
on Serial connectivity (RS-232)
• How?
Comparing IoT Architecture
Two best known IoT Standard architecture
• oneM2M IoT Standardized Architecture
• IoT World Forum (IoTWF) Standardized
Architecture
oneM2M IoT Standardized
Architecture
• In an effort to standardize the rapidly growing field
of machine-to-machine (M2M) communications,
the European Telecommunications Standards
Institute (ETSI) created the M2M Technical
Committee in 2008.
• Goal: create a common architecture that would
help accelerate the adoption of M2M applications
and devices and expanded to include IoT.
Main Elements of oneM2M IoT
Architecture
• Application Entities (AEs) reside within individual
device and sensor applications.
• Provides a standardized interface to manage and
interact with applications.
• Common Services Entities (CSEs) resides between
the applications layer and the in the network layer.
• The network layer ensures that devices and sensors
and applications are able to function in a network-
agnostic manner
Main Elements of oneM2M IoT
Architecture (Top down)
IoTWF Standardized Architecture
• In 2014 the IoTWF architectural committee (led by

Cisco, IBM, Rockwell Automation, and others)
published a seven-layer IoT architectural reference
model
1. Physical Device and Controllers
2. Connectivity
3. Edge Computing
4. Data Accumulation
5. Data Abstraction
6. Application
7. Collaboration and Processes
IoTWF Architecture layers
IT and OT (IoT) responsibilities
in the IoTWF Architecture
Additional IoT Architecture
• Purdue Model for Control Hierarchy
• Industrial Internet Reference Architectgure (IIRA) by
Industrial Internet Consortium (IIC)
• Internet of Things-Architecture (IoT-A)
Simplified IoT Architecture
To be Continued..
Homework
Jika direlasikan dengan sebuah sistem IoT meteran listrik (Smart Meter),
Jelaskan tanggungjawab dan contoh tidakan pada setiap layer yang ada
pada arsitektur IoTWF untuk sistem IoT Smart meter tersebut.
Layer Tanggungjawab/Tindakan
Part2
• Simplified IoT Architechture
• The Core IoT Functional Stack
• IoT Data Management and
Compute Stack
Simplified IoT Architecture
• The framework grouped all layers in two:
• Core IoT Functional Stack
Expanded Architecture
Core IoT Functional Stack
Layer 1: Things: Sensor and
Actuator Layers
• The object or “thing” that needs to be connected
(smart objects)
• Smart Object (things) classification parameters:
• Battery-powered or power-connected
• Mobile or static
• Low or high reporting frequency
• Simple or rich data
• Report range
• Object density per cell
Example of sensor classification
based on mobility and Throughput
Layer 2: Communication
Network Layer
• This layer responsible to connect the “things” after
its form factors, classifications determined.
• Sub layers:
• Access Network
• Gateway and Backhaul
• Network Transport
• IoT Network Management
Access Network
• Provides direct connectivity between the “things” and the network
Access Technologies and Distance

• 4 factors to determine access technology
• Range
• Expected range between the things and the information
collector
• Throughput
• The amount of data and transfer frequency
• Power Consumption
• Battery powered? High transfer rate require more power
• Cost
• Different technology has different implementation cost
Range vs Throughput for WHAN to WLAN
Range vs cost vs power vs bandwidth
Access Network Topology
• Basic topologies
• Point to Point
• Point to Multipoint
• Extended Topology
Gateway and Backhaul
• Data collected from a smart object through an access technology may
need to be forwarded to a central station where data in processed
• Gateway or backhaul technology provides higher bandwidth and
reliable connection to transfer the collected data
Network Transport Sublayer
• The protocols that enable communication between
devices with chosen access technology
• The protocols need to be open and standard based
• IP (TCP/UDP) is the most common and most used
protocol
IoT Network Management
Sublayer
• Network transport protocols enable data transfer,
IoT Network Management Enable Control
• IoT specific network protocols:
• Constrained Application Protocol (CoAP)
• Message Queue Telemetry Transport (MQTT)
Layer 3: Applications and
Analytics Layer
• Application
• Analytics application
• Control application
• Analytics
• Data analytics
• Network analytics
IoT Data Management and
Compute Stack
• How to handle massive amount of data generated by the
“things”?
• Connect “things” to central cloud?
• Requirement of IoT Data management
• Minimizing Latency
• Milliseconds matter for many types of industrial systems
• Analyzing data close to the device is often required solution
• Conserving network bandwidth
• Offshore oil rigs generate 500 GB of data weekly, Commercial jets generate 10
TB for every 30 minutes of flight. It is not practical to use centralized cloud
• Increasing local efficiency
• Certain types of data need to be analyzed locally (e.g. environment condition)
IoT Data management and
Compute Stack
• Hierarchical and distributed data management
• Fog Computing
• Place computing resources as close to the ground – IoT
devices – as possible
• Edge Computing (mist computing)
• Extend the fog computing by placing the computing
resources into the device itself
• Cloud Computing
• Core datacenter and data processing
Hierarchy of Edge, Fog, and
Cloud
Summary
• The Core IoT Functional Stack
• IoT sensor and actuators layer
• Networking components layer
• Application and analytics layer
• Deals with how and where data is filtered, aggregated,
stored and analyzed.
• Differ to traditional IT models, data management in IoT
architecture is distributed as close to the edge as
possible.

Iot Network Architecture and Design: Mms Institut Teknologi Del

Uploaded by

Copyright:

Available Formats

Iot Network Architecture and Design: Mms Institut Teknologi Del

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Iot Network Architecture and Design: Mms Institut Teknologi Del

Uploaded by

Copyright:

Available Formats

IoT Network Architecture and

• Most IoT sensors are designed for a single

• Data is generated by millions of nodes

• Different from traditional IT Network devices, IoT

• In 2014 the IoTWF architectural committee (led by

Access Technologies and Distance

You might also like