Internet of things a survey on enabling technologies, protocols and applications
1. Internet of Things: A Survey on Enabling
Technologies, Protocols and Applications
ECE-653 Advance Computer Networking
Course Seminar
Mustafa Sadiq Aljumaily
2. Agenda
- What is IoT?
- Market Opportunities for IoT.
- IoT Architecture.
- IoT Elements.
- IoT Common Standards.
- QoS Criteria, IoT Challenges and Future Directions.
- Big Data Analytics, Cloud and Fog Computing in Support of the IoT.
- The Need for Better Horizontal Integration Between Application Layer
Protocols.
- Application User-Case
3. Internet of Things in Brief
- The IoT is enabled by the latest developments in RFID, smart sensors,
communication technologies and Internet protocols.
- The basic premise is to have smart sensors collaborate directly without human
involvement to deliver a new class of applications.
- IoT is expected to bridge diverse technologies to enable new applications by
connecting physical objects together in support of intelligent decision making.
- IoT applications are ranging from smart homes, transportation, healthcare,
industrial automation, and emergency response to natural and manmade disasters
where human decision making is difficult.
- IoT can make the devices connected to the Internet smart by using ubiquitous
and pervasive computing, embedded devices, communication technologies, sensor
networks, Internet protocols and applications.
- New devices, protocols, and architecture standardizations need to be developed
for the IoT to achieve such purpose.
4. Vertical and Horizontal Market of IoT
every domain specific application is
interacting with domain independent
services, whereas in each domain
sensors and actuators communicate
directly with each other
5. Market Opportunity
- IoT is a great opportunity for
equipment manufacturers, ISPs, and
Application developers.
- Several Billions and Trillions of Dollars
are the expectations of growth in
IoT-based services.
- Healthcare is expected to be the
largest portion of IoT business.
- Unique opportunity for traditional
equipment and appliances to
transform their products to smart
things.
- ISPs are required to provide QoS for
mixed M2M, P2M, and P2P traffic.
6. IoT Architecture
- interconnecting billions or trillions of heterogeneous objects through the
Internet, increases the need for a flexible layered architecture.
- Multiple architectures were proposed.
- 3 layers are good for WSNs,
but not IoT.
- 5 layers are more
representative of IoT special
requirements.
- Five-layer architecture
is the most applicable model
for IoT applications.
7. IoT 5-Layers Architecture
1- Objects Layer: includes sensors and actuators to create the big data of the IoT.
2- Object Abstraction Layer: transfer data from objects layer to service management layer
through various technologies. Cloud computing and data management processes are
handled in this layer.
3- Service Management Layer: the middleware layer to pair services with requestors
based on addresses and names. Process the received data, make decisions, and deliver the
services to the network wire protocols.
4- Application Layer: provides services to customers. Smart homes, smart buildings,
transportation, industrial automation and smart healthcare.
5- Business Layer: Manages the overall IoT system activities and services. Building business
model, graphs, flowcharts, ...etc. Design, analyze, implement, evaluate, monitor, and
develop IoT related elements. Support decision making processes.
11. IoT Common Standards
- Different protocols and standards
are proposed to facilitate and
simplify application programmers
and services providers’ jobs.
- IEEE, W3C, IETF, ETSI and other
organizations are working on
developing and supporting these
standards.
- Most prominent protocols are
shown in the figure.
12. Application Protocol (CoAP)
- Web transfer protocol based on REST on top
of HTTP.
- Bound to UDP (not TCP) to be more suitable
for IoT applications.
- Two sublayers (messaging sublayer for
reliable communication, and request/response
sublayer to handle REST communications).
- utilizes four types of messages: confirmable,
non-confirmable, reset and acknowledgement.
- GET, PUT, POST and DELETE to achieve
Create, Retrieve, Update and Delete (CRUD)
operations.
- Features provided by CoAP includes resource
observation, block wise resource transport,
resource discovery, and security.
14. Application Protocol (XMPP)
- IETF Instant Messaging (IM) standard.
- achieve authentication, access control,
privacy measurement, hop-by-hop and
end-to- end encryption, and compatibility
with other protocols.
- An XML stanza represents a piece of code
that is divided into three components:
message, presence, and iq (info/query).
- The text based communication in XMPP
using XML imposes high network
overhead. One solution to this problem is
compressing XML streams using EXI
15. Application Protocol (AMQP)
AMQP defines two types of messages: bare
messages that are supplied by the sender
and annotated messages that are seen at
the receiver
16. Application Protocol (DDS)
- Publish- subscribe protocol for real
time M2M communications.
- Two layers DCPS and DLRL.
- DCPS is responsible for delivering the
information to the subscribers.
- DLRL is an optional layer serves as the
interface to the DCPS functionalities.
- Five entities involved: publisher,
datawriter, subscriber, datareader, and
topic.
18. Service Discovery Protocol (mDNS)
mDNS is an appropriate choice for
embedded Internet-based devices due
to the facts that:
a) There is no need for manual
reconfiguration or extra administration
to manage devices.
b) It is able to run without infrastructure
c) It is able to continue working if failure
of infrastructure happens.
19. Service Discovery Protocol (DNS-SD)
- The pairing function of required
services by clients using mDNS is
called DNS-based service
discovery (DNS-SD).
- The main drawback of these two
protocols is the need for caching
DNS entries especially when it
comes to resource- constrained
devices.
20. Infrastructure Protocol (RPL)
- IETF link independent routing
protocol.
- Based on IPv6.
- Destination Oriented Directed Acyclic
Graph (DODAG) is the core of RPL.
- Single root.
- Each node is aware of parent, but no
children information.
- Non-storing mode: routing towards
lower levels based on IP source
routing.
- Storing mode: downward routing
based on destination IPv6 address.
21. Infrastructure Protocol (6LowPAN)
- Communications protocol with special characteristics like limited packet size,
various address lengths, low bandwidth.
- The specification of mapping services required by the IPv6 over Low power
WPANs to maintain an IPv6 network.
- Four types of headers: (00) NO 6LoWPAN Header, (01) Dispatch Header, (10)
Mesh Addressing, and (11) Fragmentation
- NO 6LoWPAN Header: packets that do not accord to the 6LoWPAN
specification will be discarded.
- Compression of IPv6 headers or multicasting is performed by specifying
Dispatch header.
- 6LoWPAN removes a lot of IPv6 overheads in such a way that a small IPv6
datagram can be sent over a single IEEE 802.15.4 hop in the best case.
22. Infrastructure Protocol (IEEE 802.15.4)
- MAC and PHY for low-rate wireless private
area networks (LR-WPAN).
- Low data rate, low cost, high message
throughput, reliable communications.
- Handles a large number of nodes up to
65k.
- Supports 3 frequency channel bands (2.4
GHz, 915 MHz, and 868 MHz).
- Two types of network nodes (FFD and
RFD).
- FFD can be coordinator or normal node.
- RFD: very simple can only communicate
with coordinator in star topology.
23. Infrastructure Protocol (BLE)
- Also called Bluetooth smart.
- Cover up to 100m with 15 times
shorter latency than bluetooth.
- Proved efficiency in smartphones,
vehicle to vehicle
communications, and WSN.
- Allow devices to operate as
masters or slaves in a star
topology.
- Devices are communicating or
going to sleep mode.
24. Infrastructure Protocol (EPCglobal)
- EPC is Unique ID stored in RFID tag.
- Used in supply chain management.
- The underlying architecture uses
Internet-based RFID technologies along
with cheap RFID tags and readers to
share product information.
- EPCglobal Network: EPC, ID system, EPC
Middleware, Discovery Services, and EPC
Information Services.
- The ID system links the EPC identities to
a database using an EPC reader through
the middleware.
- The discovery service is a mechanism of
EPCglobal to find the required data by
the tags using the ONS.
25. Infrastructure Protocol (LTE-A)
- A set of cellular communication protocols fits well for IoT smart cities.
- Physical layer uses OFDMA where bandwidth is partitioned into smaller bands.
- Employs a multiple component carrier (MCC) spread spectrum to allow up to
five 20 MHz bands.
- The architecture of LTE-A network relies on two essential parts.
- The first one is the Core Network (CN) which controls mobile devices and deals
with IP packet flows.
- The other part is the Radio Access Network (RAN) which handles wireless
communication and radio access and establishes user plane and control plane
protocols.
- Challenges include network congestion and QoS compromising.
26. Infrastructure Protocol (Z-wave)
- Low power wireless communication protocol for Home Automation Networks
(HAN).
- Used in remote control applications in smart homes.
- Covers about 30 meters point-to-point communication.
- Operates in ISM bands around 900 MHz.
- Transmission rate up to 200Kbps.
- Optional ACK messages.
- Controller and slaves nodes.
- Routing is performed using source routing method.
Generally: there is a need for more investigation on improvements and
optimizations of routing protocols to meet the IoT requirements.
28. Security in IoT
- IoT cannot be secured by conventional security protocols which are used on the
Internet.
- security problems should be considered in all layers of the IoT from the application to
the infrastructure layers.
- Codo is a security solution for data storage that caches data for bulk encryption and
decryption.
- IEEE 802.15.4 security protocol protect the communication between two neighboring
devices.
- IPSec is mandatory security protocol for IPv6 network layer.
- Transport Layer Security (TLS) for TCP and Datagram TLS (DTLS) for UDP are used in the
transport layer.
- Application layer can either depends on lower layers for security or use some solutions
like EventGaurd, QUIP, Lithe, OASIS, and SASL to provide encryption and authentication
- Intrusion Detection Systems (IDS) for IoT are also investigated.
29. Interoperability (IEEE 1905.1)
- The IEEE 1905.1 standard was
designed for convergent digital
home networks and
heterogeneous technologies
- an abstraction layer that hides the
diversity of media access control
topologies.
- Environmental impact of IoT
devices, large scale and green
deployment are remain open.
30. QoS Criteria, IoT Challenges and Future Directions
1- Availability: must be realized in the hardware and software levels to provide anywhere
and anytime services for customers.
2- Reliability: refers to the proper working of the system based on its specification.
3- Mobility: most of the services are expected to be delivered to mobile users.
4- Performance: depends on the performance of many components as well as the
performance of the underlying technologies.
5- Management: manage the Fault, Configuration, Accounting, Performance and Security
(FCAPS) aspects of all devices.
6- Scalability: the ability to add new devices, services and functions for customers without
negatively affecting the quality of existing services.
7- Interoperability: handle a large number of heterogeneous things that belong to different
platforms.
8- Security and Privacy: the lack of common standard and architecture for the IoT security.
31. Big Data Analytics in Support of IoT
- Extract analytics and consequently knowledge, by which a business can
achieve competitive advantage.
- Apache Hadoop and SciDB are hardly strong enough for big data needs
of IoT
- Facebook has used an improved version of Hadoop to analyze billions
of messages per day and offer real-time statistics of user actions.
- IoT big data analytics service known as TSaaaS using time series data
analytics to perform pattern mining on a large amount of collected
sensor data.
32. Cloud Computing for IoT
Employing CC for the IoT is not an easy
task due to the following challenges:
- Synchronization: between different
cloud vendors.
- Standardization: (due to different
vendors).
- Balancing: between general and IoT
CC services.
- Reliability.
- Management.
- Enhancement.
Many solutions are available but none of
them is a “fit for all” solution.
33. Fog Computing in support of IoT
- Fog Computing (a.k.a. cloudlets
or edge computing) can act as a
bridge between smart devices
and large-scale cloud computing
and storage services.
- Mobile network operators are
the potential providers of fog
computing since they can offer
fog services as one of IaaS, PaaS,
or SaaS models to the enterprise
businesses by providing services
at their service network or even
cell tower.
34. BETTER HORIZONTAL INTEGRATION BETWEEN APPLICATION LAYER
PROTOCOLS- Two types of devices: resource
constrained and resource rich
devices.
- Interoperability is difficult.
- fragmentation between the
protocols utilized for
communication within and across
resource-constrained and
resource-rich devices is not
foreseen to change in the near
future.
- Many suggested solutions.
35. Application User Case (Nursing House)
Interested in collecting patient’s vital sign measurements, light sensor, and door sensor.
Implementation steps:
- Application developer can use SmartThings or BITalino sensors for collecting data.
- Utilize their APIs to build an application that sends collected data to nursing stations.
Or use:
- USB sensors with microcontrollers in SBC.
- Utilize WiFi or IEEE 802.15.4 to communicate their measurements.
- Download and install open source MQTT broker like Mosquitto.
- Open source implementation of MQTT like Eclipse Paho.
- Phidgets USB sensors to collect vital signs, light and door sensor data.
- MQTT clients ⇒ MQTT broker ⇒ MQTT server ⇒ Nursing Stations.