2. Recent Trends in
Computing
• Shared pool of configurable
computing resources
• Ubiquitous, dynamic & on-demand access
2
Homogeneous computing nodes (connected loosely or tightly) working together
Heterogeneous computing nodes distributed over a wide area to perform very large
tasks
Packaged resources available for computing and storage
4/28/2024
3. Evolution of Cloud Computing
1950s
Time- shared
mainframe
computers
1970s
Virtual
Machines by
IBM
3
1996-97
‘Cloud
Computing'
2002
Amazo
n Web
Service
s
(AWS)
1969 1990s 1999 2006
ARPAN
ET
Expansion Salesforce. Amaz
on
of the com EC2
Internet.
Inception
of VPNs.
2008
Google
App
Engine
/ Micorsoft
Azure
4/28/2024
4. Cloud Computing
4
“Cloud computing is a model for
enabling convenient, on-demand
network access to a shared pool of
configurable computing resources (e.g.,
network infrastructures, servers,
storage, applications, etc.)” – NIST
Source: P Mell & T Grance, “A NIST Notional Definition of Cloud Computing”, version 15,2009.
4/28/2024
5. NISTVisual Model of Cloud
Computing
Software-as-a-
Service
(SaaS)
Platform-as-a-
Service
(PaaS)
Infrastructure
-as-a-
Service
(IaaS)
Essential
Characteristics Service Models
Deployment Models
Broad
Network
Access
Rapid Elasticity
Measured Services
On-demand
Self- services
Resource Pooling
Public
Private
Hybrid
Community
5
4/28/2024
6. BusinessAdvantages
6
Nearly zero cost for upfront infrastructure investment
Real-time Infrastructure availability
More efficient resource utilization
Usage-based costing
Reduced time to market
4/28/2024
7. General Characteristics
• Improved agility in resource provisioning.
• Ubiquitous– independent of device or location
• Multitenancy– sharing of resources and costs across a large pool of
users
• Dynamic load balancing
• Highly reliable and scalable
• Low cost and low maintenance
• Improved security and access
4/28/2024 7
8. Essential Characteristics
8
Broad network access
Cloud resources should be available over the network
Should support standard mechanisms for information
retrieval using traditional interfaces
Supported clients: heterogeneous thin or thick client
platforms (e.g., mobile phones, laptops, and PDAs)
4/28/2024
9. Essential Characteristics
9
Rapid elasticity
Cloud resource allocation should be rapid, elastic and
automatic
Dynamic allocation/release facility for scale-out and
scale-in
Consumers should feel infinite resources
Facility for add/remove of quantity should be there
4/28/2024
10. Essential Characteristics
10
Measured service
Resource usage should be recorded and monitored
Facility to dynamically control and optimize the resource
usage
This facility should be transparent between the service
provider and consumer.
4/28/2024
11. Essential Characteristics
On-demand self-service
• Provide server time and network storage to users
automatically
• This facility should be available as a self-service
4/28/2024 11
12. Essential Characteristics
12
Resource pooling
• Automatically pool the whole available resources
• Serve multiple end-users using a multi-tenant model
• Resources should be allocated according to user’s
demand
4/28/2024
13. Components of Cloud Computing
13
Clients
Services
Applications
Platform
Storage
Infrastructure
Source:
Wikipedia
•Clients /end-users: Thick, Thin, Mobile
•Services: Products & solutions (Identity,
Mapping, Search, etc.)
•Applications: Web apps, SaaS, etc.
•Platform: Apps/Web hosting using PaaS
•Storage: Database, Data-Storage-as-a-Service
(DSaaS)
•Infrastructure: Virtualization, IaaS, EC2
4/28/2024
15. Software- as-a- Service(SaaS)
15
•Facility to execute service provider’s applications at user’s end
•Applications are available as ‘services’
•Services can be accessed via different types of client devices (e.g.
web browser, app)
•End-users do not posses the control of the cloud infrastructure
Examples: Google Apps, Salesforce, Learn.com.
4/28/2024
16. Platform-as-a-Service (PaaS)
16
Facility for the consumer to execute consumer-created or acquired
applications onto cloud infrastructure
Support for deployment of such applications
The user does not control the cloud infrastructure
User can control the deployed applications using given
configurations
Examples: Windows Azure, Google App Engine
4/28/2024
17. Infrastructure-as-a- Service (IaaS)
17
Facility to access computing resources such as network, storage,
and operating system
User can deploy, execute and control any software (Operating
systems and other applications)
In some case, the user can control selected networking components
(e.g., host firewalls).
Examples: Amazon EC2, GoGrid, iland, Rackspace Cloud Servers.
4/28/2024
18. Deployment Models
Public cloud
Private cloud
Hybrid cloud
Others:
Community cloud
Distributed cloud
Multi-cloud
Inter-cloud
Private/
Internal
Off-
premise
cloud
service
Public/
Hosted
On-
premise
cloud
service
Hybrid
18
4/28/2024
19. Public Cloud
• Cloud set-up for the use of any person or industry
• Typically owned by an organization who offers the cloud service
• Examples: Amazon Web Service (AWS), Google Compute Engine,
Microsoft Azure
Advantages:
• Easy to set-up at low cost, as provider covers the hardware, application
and bandwidth costs
• Scalability to meet needs.
• Pay-per-use ensures that from user’s perspective no resources wasted.
4/28/2024 19
20. Private Cloud
• Cloud set-up functioned only for a single organization
• Typically managed by the organization itself (on-premises) or a third
party (off-premises)
Advantages:
• Total control over the system and data
• Minimum security concerns
Disadvantages:
Regular maintenance
4/28/2024 20
21. Public Cloud vs Private Cloud
Public Cloud Private Cloud
Virtualized
resources
Publicly shared Privately shared
Customer types Multiple Limited
Connectivity Over Internet Over Internet/private
network
Security Low High
4/28/2024 21
22. Hybrid Cloud
• Cloud set-up constructed by two or more unique
cloud set-up (private, community, or public).
• Pooled together by standardized tools.
• Supports data and application portability
(e.g., facility for load-balancing between clouds).
• Provides multiple deployment.
4/28/2024 22
23. Other Types of Cloud
Community cloud
• Shared set-up between several organizations having common
concerns (security, compliance, jurisdiction, etc.)
• Managed by internally or by third party
Distributed Cloud
• Collection of scattered set of computing devices in different
locations, however, connected to a single network
• Two types –Public-resource Computing and Volunteer Cloud
4/28/2024 23
24. Other Types of Cloud
Multi-cloud
• Multiple cloud computing services offered via single
heterogeneous architecture
• Increases fault-tolerance and flexibility
Inter-cloud
• Unified global ‘cloud of clouds 'based on the Internet
• Supports interoperability between cloud service providers
4/28/2024 24
25. Comparison of Different Deployment
Models
25
On-premise Off-premise
Dedicated
Access
Private cloud Hosted private cloud
Shared
Access
Community cloud Public cloud
4/28/2024
28. Comparison of Different Service
Models
Traditional
Application
s Data
Runtime
Middleware
OS
Virtualizatio
n Servers
Storage
Networking
By
Service
Provider
By
User
Application
s Data
Runtime
Middleware
OS
Virtualizatio
n Servers
Storage
Networking
By
Service
Provider
Application
s Data
Runtime
Middleware
OS
Virtualizatio
n Servers
Storage
Networking
By
User
By
Service
Provider
IaaS PaaS SaaS
Applications
Data
Runtime
Middleware
OS
Virtualization
Servers
Storage
Networking
By
User
3
4/28/2024
31. Why IaaS?
6
New businesses can operate without investing on computer
hardware.
Scalable for rapidly growing businesses. (Organizations that
experience huge success immediately)
Suitable for serving fluctuating computing demands. (Ex. Flipkart,
Amazon during festival seasons)
Suitable for new business model trials.
Helps in minimizing the capital expenses. (entrepreneurs
starting on a shoestring budget)
4/28/2024
32. Essential Characteristics
7
Scalability and elasticity:
Dynamic scaling of required infrastructure resources
Large amount of resource allocation/release in a short
span of time
No variation in system performance while scale in or out
4/28/2024
33. Essential Characteristics
8
Manageability and interoperability
Clients have control of the virtualized infrastructure resources
Pre-configured facility for allocation of virtualized
resources
The virtualized resources are to be monitored for their running
status
The Usage and Billing system records the use of
infrastructure resources and accordingly calculate
payment
4/28/2024
34. Essential Characteristics
9
Availability and reliability
Stored data can be retrieved at any time without failure
The clients should be able to access the computational
resources without failure
Uninterrupted facility for computation and
communication
4/28/2024
35. Essential Characteristics
10
Performance and optimization
High utilization of physical resources among different
clients
To enable high computing power with the large
pool of physical resources using parallel processing
To optimize the deployment of physical resources
by dynamic configuration of virtual infrastructure
resources
4/28/2024
36. Essential Characteristics
11
Accessibility and portability
• Facility to client have various tasks – control, manage
and access infrastructure resources
• To facilitate easy reallocation and duplication of
allocated infrastructure resources
4/28/2024
38. IaaS – Challenges and Limitations
13
Sometimes the regulatory approval does not allow outsourcing
the storage and processing of sensitive data.(Ex.: Medical
records)
Network latency may degrade the level of expected
performance
4/28/2024
39. IaaS – Challenges and Limitations
(contd.)
14
Users may require automated decision making of
job scheduling to available resources
Seamless scaling of services independent of traffic variation
Developers have to focus on low level system details
4/28/2024
41. PaaS (contd.)
16
Facilitates development and managing applications
without the complexity of maintaining the underlying
infrastructure
Allows customers to rent virtualized servers and
associated services
Provide electronic scaling of the user’s deployed application
4/28/2024
42. Features of PaaS Offering
17
Operating system
Service side scripting environment
Database management system
Server Software
Support
Storage
Network access
Tools for design and development
Hosting
4/28/2024
43. PaaS Working Model
18
•Allows users to create software applications using offered tools
•Provides preconfigured features that customers can subscribe
•Support available for managing the infrastructure and applications
for customers
•Services are regularly updated with new features
4/28/2024
44. Business Advantages
19
Facility for accessing key middleware services without
worrying about the underlying complexities of managing
individual hardware and software elements.
Ease of access for the development and deployment tools.
Freedom from managing development and deployment tools
individually.
4/28/2024
45. Software- as-a- Service(SaaS)
20
“Software as a Service provides you with a completed product that is
run and managed by the service provider. In most cases, people
referring to Software as a Service are referring to end-user
applications.” – Amazon
Source: https://aws.amazon.com/types-of-cloud-computing/
SaaS is a simplified model of software delivery over Internet.
Operation, maintenance and technical support is provided by the
service provider.
Typically offered via web browser working as a thin-client.
Supports a fully pay-as-you-go model.
Source: Software Services for e-Business and e-Society: Proceedings of 9th IFIP WG 6.1 Conference on e-Business, e-Services and e-Society,
I3E 2009, Nancy, France, September 23-25, 2009.
4/28/2024
46. SaaS (contd.)
21
Remote access of software via Internet where web-
browser acts as a thin-client
Facility for access and control of commercial
software via Internet
delivery in a one-to-many model
4/28/2024
47. Advantages
22
Traditional Software SaaS
Customers install, manage &
maintain
Customers uses over the Internet
Runs on individual
organization on dedicated
instantiation
Runs on multiple
customers
simultaneously
Cross platform support required No concerns for cross platform
support
Less frequent version
updates & purchased
separately
More frequent updates for
enhanced user satisfaction
Separate costs incurred for
upgrades
No separate cost
Vulnerable to software piracy Less vulnerable to software piracy
4/28/2024
48. SaaS Architecture
23
Scalability
To maximize application concurrency
•To optimize the shared pool of resources such as threads and network
connections
Multi-tenancy
•Important architectural shift from designing isolated, single-tenant
applications
•Ability to accommodate users from multiple companies at the same time
•Transparency to all the users
•Maximize the sharing of resources across tenants while distinguishing user’s
individual data
4/28/2024
49. SaaS Architecture (contd.)
24
Configurability
To facilitate parallel allocation of a single application on a
single server to several users.
To customize the application for one customer will change
the application for other customers as well.
Separate data space for different users
4/28/2024
50. Limitations of SaaS
25
Centralized control
Switching cost
Limited flexibility
Data security and privacy
4/28/2024
52. Introduction
2
Simulation tools provide reliable, scalable and repeatable
environment for performance evaluatio
The simulator facilitate pre-deployment tests of services
As the demand of cloud computing is growing everyday, the
simulators and technologies are needed to be studied
4/28/2024
53. Introduction (contd.)
3
Cloud simulators allow customers to
Evaluate the services
Testing at no cost
Enable repeatable evaluation
Control the environment
Pre-detection of issues affecting
performance
Design of countermeasures
4/28/2024
55. CloudSim
5
A simulation framework
Models cloud computing environments– Data Center,
VM, applications, users, network topology.
Written on Java-based environment.
Allows to examine the performance of application services.
Dynamic addition/removal of resources during simulation.
Developed at CLOUDS Lab. of University of Melbourne.
4/28/2024
56. Advantages of CloudSim
6
Time effectiveness: Cloud-based application
implementation in
Minimum time
Minimum effort
Flexibility and applicability:
Support for diverse cloud environments
Enables modelling of application services in any
environment
4/28/2024
57. Features of CloudSim
7
Various cloud computing data centers
Different data center network topologies
Message-passing applications
Virtualization of server hosts
Allocation of virtual machins (VMs)
User defined policies for allocation of host resources to
VMs
Energy-aware computational resources
Dynamic addition/removal of simulation components
Stop and resume of simulation
4/28/2024
58. CloudSim Architecture
User Code: Top most layer
Presents different machine and application specifications
CloudSim: Middle layer
Provides cloud environment
Enables modelling and simulation
Core Simulation Engine: Bottom most layer
Event scheduling
Entity creation
Interaction between components
Clock management
4/28/2024 58
59. Top Layer: User Code
Basic entities:
Users
Physical Machines
Virtual Machines
Applications &
services
Scheduling policies
Fig: Functionalities at top
layer
Simulation
Specificatio
n
Schedulin
g
Policy
Use
r
Cod
e
Cloud
Scenari
o
User
Requireme
nt
Application
Configuratio
n
User Broker
Data
Center
Broker
9
4/28/2024
60. Middle Layer:CloudSim
10
Creation and simulation of
Dedicated management interfaces
Memory, storage, bandwidth and VMs
Helps in solving issues like
Hosts provisioning to VMs
Application execution management
Dynamic system state monitoring
Allows a cloud service provider to
Implement customized strategies
Evaluating the efficiency of different policies in VM
provisioning
4/28/2024
62. CloudAnalyst
12
Simulation tool designed based on CloudSim
Provides GUI
Supports geographically distributed large-scale Cloud
applications
The purpose is to study the behavior of such applications under
various deployment configurations
4/28/2024
63. Features of CloudAnalyst
13
Easy to use due to Graphical User Interface (GUI)
High level of configurability
Flexibility of adding components
Repeatability of experiments
Graphical output (e.g. charts, tables)
Easy to extend (Java Swing) and uses blended technology
4/28/2024
64. CloudAnalyst Design
Fig: CloudAnalyst
Architecture
CloudSim
Toolkit
CloudSi
m
Extension
s
GUI
CloudAnaly
st
14
Main components
• GUI Package: Front end
• Simulation: Create, execute, hold
• UserBase: User traffic generation
• DataCenterController: Events of data center
• Internet: Internetworking & routing
• InternetCharacteristics: Properties of Internet
(delay, Bandwidth, throughput, etc.)
• VmLoadBalancer: Policies for load balancing
• CloudAppServiceBroker: Entities for routing
between
UserBase & data center.
4/28/2024
65. GreenCloud
15
Why:
The computing capacity has increased the cost and operational expenses
of data centers
Energy consumption by data center is the major factor driving the
operational expense
What:
Operational cost is the energy utilized by computing and
communication units within a data center
How:
GreenCloud monitors the energy consumption of servers, switches, etc.
Developed as an extension of a packet-level network simulator NS2
4/28/2024
66. Features of GreenCloud
GU
I
iUesnedrl-yfr
Open source
Facility for monitoring energy consumption of
network & devices
Srtusppo simulation of cloud network components
Supports monitoring of energy consumption of
individual components
Enables improved power management schemes
Dynamic management and configuration of devices
4/28/2024 66
67. Open Source and Commercial
Clouds
17
Open Source Clouds Commercial Clouds
Examples OpenStack,
CloudStack,
Eucalyptus
Amazon Web Services
(AWS), Microsoft Azure,
Google App Engine
Facility Mostly offers IaaS IaaS, PaaS, SaaS
Services on
subscription
Security Implemented by user Implemented by service
provider
Type Private/On-premise Public/Off-premise/Hosted-
private
4/28/2024
68. OpenStack
Collection of open source
technologies
Managed by the OpenStack
Foundation
Supports vastly scalable cloud system
Preconfigured software suit
Different services available for users
Considered Infrastructure as a
Service (IaaS).
Ease of use: add new instances
quickly to run other cloud
components
Provides a platform to create software
applications
Developed software applications
can be used by the end users
VM
Container
Storage
Dashboard
GUI
Monitorin
g Tools
Apps Apps
Common Network
18
User’
s
Apps
4/28/2024
69. Microsoft Azure
20
Previously Windows Azure
Supports Iaas and PaaS
Supports extensive set of services to quickly create, deploy
and manage applications
Many programming languages and frameworks are supported
Available across a worldwide Microsoft-managed datacenters
4/28/2024
70. Azure Services
21
Compute
Mrv
iocbeislese
Srvtoicreasgese
Danaat
agemment
Messaging
r
M
v
i
e
c
d
e
i
s
ase
Content Delivery Network
(CDN)
Developer
Management
Machine Learning
4/28/2024
71. Azure as PaaS (Platform as a Service )
22
Platform is provided to clients to develop and deploy software
Clients focus on application development rather than worry
about hardware and infrastructure
Low Cost
less vulnerable to security attacks
Ease to move on to new tools
Solves the issues related to most of the operating systems,
servers and networking.
Source URL : https://azure.microsoft.com/en-in/overview/what-is-paas/
4/28/2024
72. Azure as IaaS (Infrastructure as a
Service )
23
Offers total control of the OS and application
stack
Features to access, manage and monitor the
data centers
Ideal for the application where complete
control is required
Facility for loading of custom configurations
4/28/2024
73. Amazon Elastic Compute Cloud (EC2)
24
A web service for users to launch and manage
server instances in Amazon’s data centers
Provides various APIs, tools and utilities
Facilitate dynamic computation scaling in the Amazon
Web Services (AWS) cloud
Supports pay-per-use billing rather than making large and
expensive hardware purchases
Source: amazon web services Website https://aws.amazon.com/ec2/
4/28/2024
74. Amazon EC2 Instances
25
Virtual computing environments
Instance templates of different configurations – CPU, memory,
storage, networking capacity
Dynamic instance allocation by AWS according to user demand
Instance types
General purpose: T2, M4, M3
Compute optimized: C4, C3
Memory optimized: X1, R4, R3
Accelerated computing instances: P2, G2, F1
4/28/2024
75. Features of Amazon EC2
26
Operating system:
Supports all OS types
Custom distribution: Amazon Linux AMI/Amazon Machine Images
Persistent storage:
Temporary: Local ‘Instance Store’
Amazon Elastic Block Store (EBS)
Simple Storage Service (S3)
Automated scaling: Rule based / Schedule based
Different “availability zones” in data centers increases fault-tolerance
4/28/2024
76. Features of Amazon EC2
7
Firewall Rules/Security Groups: Only predefined
protocols, ports, and source IP ranges reach the instances
Elastic IP address: Mapping between IP and any VM of user
Amazon CloudWatch: CPU, disk, network resource
utilization monitoring
Enhanced security for instances using public-private key
pair
Virtual private clouds (VPCs):
Logically separate from the rest of the AWS cloud
Optionally connected to user’s own network
4/28/2024
82. Components contd.
82
Nova
Compute service
Where you launce yourinstances
Glance
Image service
Discovering, registering, retrieving theVM
Snapshots
4/28/2024 82
83. Components contd.
83
Swift
Object storage
Helps in storing data safely,cheaply and efficiently
Neutron
Provides networkingservice
Enables the other services to communicate with each other
Make your ownnetwork
4/28/2024 83
84. Components contd.
84
Cinder
Block storage
Virtualizes the management of blockservice
Heat
Orchestration
Ceilometer
Billing
What service you areusing
How long are youusing
4/28/2024 84
85. Installation
85
Canbe installedmanuallyor using scripts likeDevstack
We will usedevstack
Steps:
Installgit(sudo apt‐getinstallgit)
Clone devstack (git clone
https://git.openstack.org/openstack‐ dev/devstack)
Go todevstackdirectory(cddevstack)
4/28/2024 85
86. Installation contd.
86
Open local.conffileand pastethefollowingand savethefile
ADMIN_PASSWORD=<YOUR PASSWORD>
DATABASE_PASSWORD =<YOUR
PASSWORD> RABBIT_PASSWORD =<YOUR
PASSWORD> SERVICE_PASSWORD =<YOUR
PASSWORD>
HOST_IP=<the IP of yourPC>
Run the stack.sh file(./stack.sh)
Foruninstallation,go todevstackdirectoryand run unstack.shfile
4/28/2024 86
90. Introduction
It is not mere integration of sensors and cloud computing
It is not only “dumping the sensor data into cloud”
Cloud
Cloud
2
91. Wireless Sensor Networks
(WSNs): Recap
3
Contain sensor nodes which sense some physical phenomena
from the environment.
Transmit the sensed data (through wireless communication) to a
centralized unit, commonly known as Sinknode.
The communication between Sink node and other sensor
nodes in the network may besingle/multi‐hop.
Sink node further processdata.
92. Wireless Sensor Networks (WSNs):
Recap
Sensing unit
Processing unit
Communication unit
Major Components ofa
Sensor Node
Sink
Wireless SensorNetworks
4
Applications
T
argetTracking
Wildlife Monitoring
Healthcare
Industrial Applications
Smart Home
Smart City
Agriculture
…
93. Cloud Computing: Recap
5
An architecture which provides on‐demand
computingresources
Advantages
Elasticity:Scaling up/down
Pay‐per‐use: Payment for the resource as per
requirement
Self Service: Resource can be accessed byself
96. Virtualization Concept
8
One computer host appears as many computers‐concept of Virtual
Machine (VM)
Improve ITthroughput and costs by using physical resources as a pool
from which virtual resources can beallocated.
Benefit
Sharing of resources: Same resource can be shared, in turn cost
reduction
Encapsulation:A complete computingenvironment
Independence: Runs independently of underlyinghardware
Portability: VM Migration
97. Limitations of WSNs
Procurement
Price
Rightvendor
Typesof sensor integratedwithit
Deployment
Right way ofdeployment
Right placeofdeployment
Maintenance
Post deploymentmaintenance
Batterylifetime
Change of Requirement
An example
Today Tomorrow
Agriculture Smart Home
Result: Change in Sensor type,deployment
area, topology design, and manymore….
9
98. Sensor-Cloud: Introduction
10
Not only the mere integration of cloud computing and sensor
networks, but sensor‐cloud is more thanthat
Concept of virtualization of sensornode
Pay‐per‐use
One sensor node/network appears asmany
A stratum between sensor nodes and end‐users
99. Difference with WSN
Virtualization
WSN user
Aggregated
data
Dedicated to
a single user
Multiple
applications/ users
11
Serves
multiple
applications
Sensor‐cloud
infrastructur
e
WSN Sensor‐Cloud
Source: S. Misra; S. Chatterjee; M. S. Obaidat, "On Theoretical Modeling of Sensor Cloud: A Paradigm Shift From Wireless Sensor Network," in IEEE
Systems Journal , vol.PP, no.99, pp.1-10
100. Difference with WSN (Contd.)
Source: S. Misra; S. Chatterjee; M. S. Obaidat, "On Theoretical Modeling of Sensor Cloud: A Paradigm Shift From Wireless Sensor Network," in IEEE
Systems Journal , vol.PP, no.99, pp.1-10
Actors and Roles
Attributes WSN Sensor Cloud
Ownership WSN‐user Sensor‐owner
Deployment WSN‐user Sensor‐owner
Redeployment WSN‐user SCSP
Maintenances WSN‐user SCSP
Overhead WSN‐user SCSP
Usage WSN‐user End‐user
12
101. Actors in Sensor-cloud
13
End‐users
EnjoySe‐aaS through applicationsas per therequirements.
Unknown about what and which physical sensor is/areallocated to
serve the application
Sensor‐owner
Plays a role from businessperspective.
They purchase physical sensor devices, deployed over different
geographical locations, and lend these devices to thesensor‐cloud
Sensor‐Cloud ServiceProvider(SCSP)
A businessactor.
SCSPchargespricefromtheend‐usersas per theirusage of Se‐aaS.
103. Sensor-cloud: View
Real
View
Source: S. Misra; S. Chatterjee; M. S. Obaidat, "On Theoretical Modeling of Sensor Cloud: A Paradigm Shift From Wireless Sensor Network," in IEEE
Systems Journal , vol.PP, no.99, pp.1-10
Application
1
Application
2
User
organizatio
n
Browser
Interfac
e
Data feed
Data feed Sensed
informatio
n
T
emplate
display
Templatespecification
User organization view
User Login Xml
interpretation
Dynamic Scaling
On‐demandphysical
sensor scheduling
Vast data storage and
specializedprocessing
Energy
management,QoS
Application
specific real‐time
data aggregation
Web
portal
Template
specificati
on
Sensed
data
Heterogeneous
pool ofphysical
sensors
Interaction
with physical
sensor
On‐deman
d sensor
data
104. Work Flow of Sensor-Clou d
Compatible
sensor
scheduling,
allocation,
deallocation
Operations
request Create virtual
sensor instance
User
organization
SensorML
interpretor
Virtual Sensor
Manager
Virtual Sensor
Controller
Resource
Manger
Manage
operations
Response
Response
Data request
XML template Decode
Physical sensordefinition,
Virtual sensor
Group definition
Client information
Metadata
Templates
Sensor
resourcepool
(WSN)
Data retrieval
Data
aggregation
Data
provisioning
Delete
virtual sensor
Release
instance
Release
resource
Source: S. Misra; S. Chatterjee; M. S.
Obaidat, "On Theoretical Modeling of
Sensor Cloud: A Paradigm Shift From
Wireless Sensor Network," in IEEE
Systems Journal , vol.PP, no.99,pp.1-10
16
105. Case Study: Target Tracking
17
“We consider a WSN‐based target tracking application, in which a WSN owner
refuses to share the sensed information with an external body, even in exchange
of money. Consequently, any organization that wishes to detect intrusion within
a particular zone has to deploy its own WSN. This leads to a long‐term
investment due to costly network setup and maintenance overheads. However,
in a sensor‐cloud environment, the same organization can use the same tracking
applicationand stillget the service without actually owning theWSN”
Source: S. Misra; S. Chatterjee; M. S. Obaidat, "On Theoretical Modeling of Sensor Cloud: A Paradigm Shift From Wireless Sensor Network," in IEEE
Systems Journal , vol.PP, no.99, pp.1-10
106. Management Issues in Sensor-Cloud
2
Optimal Composition of virtual sensornodes
Data Caching
Optimal Pricing
107. Optimal Composition of Virtual Sensor
3
Source: S. Chatterjee and S. Misra, “Dynamic Optimal Composition of a Virtual Sensor for Efficient Virtualization Within
Sensor‐cloud”, IEEE ICC2015.
108. Introduction
4
Efficientvirtualization of the physical sensornodes
An optimal composition ofVSs
Consider same geographic region:CoV‐I
Spanning across multiple regions:CoV‐II
Source: S. Chatterjee and S. Misra, “Dynamic Optimal Composition of a Virtual Sensor for Efficient Virtualization Within
Sensor‐cloud”, IEEE ICC2015.
109. Why Composition of Virtual
Sensor?
Introduction to Internet of
Things
5
Resource‐constrained sensor nodes
Dynamic change in sensorconditions
The composition of virtual sensors arenon‐traditional
Source: S. Chatterjee and S. Misra, “Dynamic Optimal Composition of a Virtual Sensor for Efficient Virtualization Within
Sensor‐cloud”, IEEE ICC2015.
110. CoV-I: Formation of Virtual Sensor
Optimal formation of Virtual
sensor nodes
geographical
Sensor (VS)
Homogeneous
within same
boundary
V
S
6
Source: S. Chatterjee and S. Misra, "Optimal composition of a virtual sensor for
efficient virtualization within sensor‐cloud," 2015 IEEE International Conference
onCommunications (ICC), London, 2015, pp.448‐453
111. CoV-II: Formation of Virtual Sensor
Group
of Virtual Sensor
Formation
Group(VSG)
Heterogeneous physical sensor
nodes across different
geographical locations
VS
1
7
VS
2
VS
3
VS
G
Source: S. Chatterjee and S. Misra, "Optimal composition of a virtual sensor for
efficient virtualization within sensor‐cloud," 2015 IEEE International Conference
onCommunications (ICC), London, 2015, pp.448‐453
112. Performance
Source: S. Chatterjee and S. Misra, "Optimal composition of a virtual sensor for efficient virtualization within sensor‐cloud," 2015 IEEE InternationalConference on
Communications (ICC), London, 2015, pp.448‐453
Introduction to Internet of
Things
8
113. Dynamic and Adaptive Data
Caching Mechanism
9
Source: S. Chatterjee, S. Misra, “Dynamic and Adaptive Data Caching Mechanism for Virtualization within Sensor‐Cloud”,IEEE ANTS
2014.
114. Introduction
10
Introduces internal and external cachingmechanisms
Ensures efficiency in resourceutilization
Flexiblewith the variedrateof change of the physical environment
Source: S. Chatterjee, S. Misra, “Dynamic and Adaptive Data Caching Mechanism for Virtualization within Sensor‐Cloud”,IEEE ANTS
2014.
115. Why Caching in Sensor-Cloud?
11
End‐users requestforthesensed informationthroughaWeb‐interface
Allocationofphysical sensor nodesandvirtualization takesplace
Physicalsensor nodescontinuously sense andtransmit datatosensor‐cloud
Source: S. Chatterjee, S. Misra, “Dynamic and Adaptive Data Caching Mechanism for Virtualization within Sensor‐Cloud”,IEEE ANTS
2014.
116. Why Caching in Sensor-Cloud?
(Contd.)
12
Practically, in some cases, the change in environmental
conditionare significantly slow
Due to the slow change in environment, the sensed data of physical
sensors unaltered
In such a situation, unnecessary sensing causes energy
consumption
Source: S. Chatterjee, S. Misra, “Dynamic and Adaptive Data Caching Mechanism for Virtualization within Sensor‐Cloud”,IEEE ANTS
2014.
117. External and Internal Caching
Mechanism
13
Internal Cache(IC)
Handles requests fromend‐user
Takesdecision whether the datashould be provided directlyto the end
useroris itrequiredtore‐cachethedatafromexternalcache
ExternalCache(EC)
After every certain intervaldataare required tore‐cache
Initially,fewdataareused tobetransmitted toIC
Source: S. Chatterjee, S. Misra, “Dynamic and Adaptive Data Caching Mechanism for Virtualization within Sensor‐Cloud”,IEEE ANTS
2014.
118. Architecture of Caching
ExistingArchitecture Cache‐enabledArchitecture
App1 Appn
Sensor‐Cloud
Resource
pooling
App2 . . .
App1 App2 Appn
. .
.
14
EC
Sensor‐Cloud
IC
Source: S. Chatterjee, S. Misra, “Dynamic and Adaptive Data Caching Mechanism for Virtualization within Sensor‐Cloud”,IEEE ANTS
2014.
119. Performance
Source: S. Chatterjee, S. Misra, “Dynamic and Adaptive Data Caching Mechanism for Virtualization within Sensor‐Cloud”,IEEE ANTS
2014.
15
120. Dynamic Optimal Pricing for Sensor-Cloud Infrastructure
Introduction to Internet of
Things
16
Source: S. Chatterjee, R. Ladia, and S. Misra, “Dynamic Optimal Pricing for Heterogeneous Service‐Oriented Architecture of
Sensor‐Cloud Infrastructure”, IEEE TSC2017.
121. Introduction
17
Existing schemes consider homogeneity of service(e.g.forIaaS, SaaS)
No scheme forSeaaS.
Theproposed pricingscheme comprises oftwocomponents:
Pricingattributed to hardware(pH)
Pricingattributed to Infrastructure(pI)
Goal of the proposed pricingscheme:
Maximizing profit ofSCSP
Maximizing profit of sensorowner
End users’satisfaction
123. Focus on
19
Maximizing the profit made bySCSP
Optimal pricingto theend‐users
End userssatisfaction
Pricingattributed to hardware(pH)
Dealswithusageofphysicalsensor nodes
Pricingattributeto infrastructure(pI)
Dealswiththepriceassociatedwithinfrastructure ofsensor‐cloud
124. References
20
Madoka Yuriyama and Takayuki Kushida ,“Sensor‐Cloud Infrastructure ‐ Physical Sensor Management with
Virtualized Sensors on Cloud Computing”, Research Report ,IBM Research ‐ Tokyo IBM Japan, Ltd.,2010
(http://domino.research.ibm.com/library/cyberdig.nsf/papers/70E4CC6AD71F2418852577670016F2DE/$F
ile
/RT0897.pdf)
S. Chatterjee, R.Ladia and S. Misra, "Dynamic Optimal Pricing for HeterogeneousService‐Oriented
Architecture of Sensor‐Cloud Infrastructure," in IEEE Transactions on Services Computing, vol. 10, no. 2,
pp. 203‐216, 2017
S. Chatterjee and S. Misra, "Optimal composition of a virtual sensor for efficient virtualization within
sensor‐ cloud," 2015 IEEE International Conference on Communications (ICC),London, 2015, pp. 448‐453
S. Misra; S. Chatterjee; M. S. Obaidat, "On Theoretical Modeling of Sensor Cloud: A Paradigm Shift
From Wireless Sensor Network," in IEEE Systems Journal ,vol.PP
,no.99,pp.1‐10
128. Introduction (contd.)
40% ofthewholeworldsdatawillcomefromsensorsaloneby2020.
90% oftheworld’sdataweregenerated onlyduringtheperiodoflasttwo
years.
2.5quintillionbytesofdatais generatedperday.
totalexpenditure on IoTdeviceswillbe$1.7Trillionby2020
129. Introduction (contd.)
the totalnumber of connected vehicles worldwide will be 250
millionsby 2020.
therewill be more than 30 billion IoTdevices
The amount of datagenerated by IoTdevices is simply huge.
131. Why Fog Computing (contd.)
Cloud
Sends data
for analysis
andstorage
Sends back
commandor
action
required
Devices
Fig.1: Present day cloud model
132. Why Fog Computing
(contd.)
DataVolume:
By2020,about50 billiondeviceswillbeonline.
Presentlybillionsofdevicesproduceexabytesofdataeveryday.
Device density is stillincreasingeveryday.
Currentcloudmodelis unabletoprocess thisamountofdata.
133. Why Fog Computing (contd.)
Cloud
Privatefirms Factories Airplane firms
Privatefirms, Factories,
airplane companiesproduces
colossus amount of data
everyday
Currentcloud modelcannot
store allthesedata
Data need to befiltered
Storingdata
134. Why Fog Computing (contd.)
Latency
Timetakenbyadatapacketforaroundtrip
Animportant aspectfor handing a time sensitive data.
Ifedgedevicessend timesensitive datatocloudforanalysisandwait for the
cloud to givea proper action,then itcanlead to many unwantedresults.
Whilehandlingtimesensitivedata,amillisecond canmakeahuge
differences.
135. Sending time‐sensitivedatato cloudfor
analysis
Latency= Tfron device to cSoud +Tdata
anaSycic
+ Tfron cSoud to device
where T = Time
Latencywill beincreased
When the action reaches thedevice,
accidentmay have already occured
Why Fog Computing (contd.)
Cloud
Analysisof
data
Appropriate
action
Sending time
sensitive datafor
analysis
136. Why Fog Computing (contd.)
Bandwidth:
Bit‐rate of dataduringtransmission
Ifallthedatagenerated byIoTdevicesaresenttocloudforstorage
and analysis, then,the trafficgenerated by these devices will be
simplygigantic.
consumes almost allthebandwidths.
Handlingthiskindoftrafficwillbesimplyaveryhardtask.
137. Why Fog Computing (contd.)
Cloud
Billionsof devicesconsumingbandwidth
If allthe devicesbecome online evenIPv6 will
not be ableto provide facilityto all thedevices
Data may be confidential which thefirms do
not want to share online
Sending datafor
analysis and
storage
Appropriate
action
138. Requirements of IoT
Reduce latency ofdata:
Appropriate actions at the right time prevents major accidents
machine failure etc.
Aminute delay while taking a decision makes a huge difference
Latency can be reduced by analyzingthe data close to the data
source
139. Requirements of IoT (contd.)
Datasecurity:
IoTdatamust besecuredandprotectedfromtheintruders.
Data are required to be monitored24x7
An appropriateactionshould betakenbeforetheattackcausesmajor
harm to thenetwork
140. Requirements of IoT (contd.)
Operation reliability:
Thedatageneratedfrom IoTdevices areused to solve real
time problem
Integrity and availabilityof the datamust be guaranteed
Unavailability and tampering of data can behazardous
141. Requirements of IoT (contd.)
Processing of data at respective suitable place:
Data can be divided into three types based on sensitivity
time sensitive data
less time sensitive data
data which are not timesensitive
Extremely time sensitive data should be analyzed very near to the data
source
Data which are not time sensitive will be analyzed in the cloud.
143. When should we use fog
If the datashould ne analyzewithfractionof second
If there are hugenumber of devices
If the devices are separated by alargegeographical distance
If the devices are needed to be subjected toextreme conditions
144. Architecture of Fog
CloudservicesareextendedtoIoTdevicesthroughfog
Fogisalayerbetween cloudandIoTdevices
many fog nodes can bepresent
Sensordataareprocessedinthefogbeforeitissenttothecloud
Reduceslatency,savebandwidth andsave thestorageofthecloud
146. Fog nodes
Characteristicsfor a fognode:
Storage ‐Togive transientstorage
Computing facility
‐ Toprocess the databefore itis sent tocloud
‐ Totakequick decisions
Network connectivity ‐Toconnectwith IoTdevices, other fognodes and
cloud
147. Fog nodes (contd.)
E.g.‐ routers, embedded servers, switches, video surveillancecameras, etc.
deployable anywhere inside thenetwork.
Each fog nodes havetheir aggregatefog node.
148. Working of Fog
Threetypes ofdata
Very time‐sensitivedata
Less time‐sensitivedata
Data which are nottime‐sensitive
Fognodes worksaccordingtothetypeofdatatheyreceive.
An IoTapplication should beinstalled toeachfognodes
149. Working of Fog (contd.)
Devices
Cloud
Nearest
FogNode
Aggregate
fog
node
Sends the summaryfor
historical analysis and
storage
Sends the summaryfor
historical analysis and
storage
Sends the summary for historical analysis and storage
Non‐time‐sensitiv
e data
Lesstime‐sensitive
data
Ingest data
If time‐sensitive
data then take
immediate action
Action
Fig :Working offog
150. Working of Fog (contd.)
Thenearestfognode ingestthedatafromthedevices.
Most time‐sensitivedata
Datawhichshould beanalyzedwithinfractionofasecond
Analyze atthe nearest nodeitself
Sends thedecisionor actiontothedevices
Sends andstoresthesummary tocloudforfutureanalysis
151. Working of Fog (contd.)
Less time‐sensitivedata
Datawhichcanbeanalyzedaftersecondsorminutes
Aresenttotheaggregatenode foranalysis
Afteranalysis,theaggregatenode send thedecisionoractiontothe device
through the nearestnode
Theaggregate node sends thesummary tocloudforstorageand future
analysis.
152. Working of Fog (contd.)
Non‐time‐sensitive data
Data whichcan be wait for hours, days,weeks
Sent to cloudfor storage and future analysis.
Those summaries fromfog nodes can be considered as
less time sensitive data.
153. Working of Fog (contd.)
Fog node
closest
to
devices
Fog aggregate nodes Cloud
Analysis duration Fraction ofsecond Seconds tominutes Hours toweeks
IoT data storage
duration
Transient Hour
,days Months toyears
Geographical
coverage
Very local Wider Global
154. Advantages of Fog
Security
Provides bettersecurity
Fognodescanuse thesame securitypolicy
Low operationcost
Dataareprocessedinthefognodesbeforesendingtocloud
Reduces the bandwidthconsumption
155. Advantages of Fog (contd.)
Reduces unwantedaccidents
Latencywill be reduceduring decisionmaking
Quick decisionmaking
Betterprivacy
Every industry can analyze their datalocally
Store confidential datain their localservers
Sendonlythosedatawhichcanbeshared tothecloud
156. Advantages of Fog (contd.)
Businessagility
Fogapplication canbeeasilydevelopedaccordingtotoolsavailable
Can be deployed anywhere weneed
Canbeprogramedaccordingtothecustomer’sneed
Supportmobility
Nodes can bemobile
Nodescanjoinandleavethenetworkanytime
157. Advantages of Fog (contd.)
Deployable in remoteplaces
Can be deployed in remoteplaces
Can be subjectedto harsh environmentalconditions
Under sea, railway tracks,vehicles, factoryflooretc
Betterdatahandling
Can operatewith lessbandwidth
Data can be analyzedlocally
Reduce the risk oflatency
158. Applications ofFog
Realtime healthanalysis
Patients withchronicillness canbemonitored inrealtime
Strokepatients
Analyze the datarealtime
During emergency,alerts the respective doctorsimmediately
Historical dataanalysis can predictfuturedangers ofthe patient
159. Applications of Fog (contd.)
Intelligencepower efficientsystem
Powerefficient
Reports detail power consumption reporteveryday
Suggest economical power usageplan
160. Applications of Fog (contd.)
Realtime railmonitoring
Fognodescanbedeployedtorailwaytracks
Realtime monitoring of the trackconditions
Forhighspeed train,sendingthedataincloudforanalysisisinefficient
Fog nodes provide fastdataanalysis
Improve safety andreliability
161. Applications of Fog (contd.)
Pipeline optimization
Gas and oils are transported throughpipelines
Real time monitoring of pressure, flow,compressor is
necessary
Terabytesof data arecreated
Sending allthis datato cloud for analysisand storage is not
efficient
Network latencyis notacceptable
Fog is asolution
162. Applications of Fog (contd.)
Realtime wind mill and turbineanalysis
Wind directionand speedanalysis can increaseoutput
Data can be monitored realtime
163. Challenges
Powerconsumption
Fog use additionnodes
Power consumption is higher than centralizedcloud
DataSecurity
Data generatingnodes aredistributed
Providingauthenticationandauthorizationsystem forthewholenodesis
not an easytask
Reliability
Maintaining dataintegrity andavailabilityformillionsofnodes is difficult
failureof a node c
e
a
c
n
t
notaff thenetwork
164. Challenges (contd.)
Faulttolerance
Failure of a node shouldbe immediately fixed
Individual failureshould not affectthe wholescenario
Realtimeanalysis
Realtime analysisisaprimaryrequirementforminimizinglatency
Dynamicanalysisanddecisionmakingreducesdangerandincreaseoutput
Monitor hugenumberofnodesisnoteasy
165. Challenges (contd.)
Programmingarchitecture
Fog nodes may bemobile
Nodes canconnectandleave thenetworkwhennecessary
Many dataprocessing frameworks are staticallyconfigured
These frameworks cannotprovide proper scalability andflexibility
178. CurrentFocusAreas
SmartHomes
Health monitoring.
Conservation of resources(e.g. electricity, water,
fuel).
Security and safety.
SmartParkingLots
Auto routing of vehicles to emptyslots.
Auto charging for servicesprovided.
Detection of vacant slots in the parkinglot.
179. CurrentFocusAreas (contd.)
Smart Vehicles
Assistance to drivers during bad weather orlow-visibility.
Detection of bad driving patterns or driving under the influence
of substances.
Auto alert generation duringcrashes.
Self diagnostics.
180. CurrentFocusAreas (contd.)
Smart Health
Low cost, portable, at-home medical diagnosis kits.
Remote check-ups and diagnosis.
On-body sensorsfor effortless and accurate healthmonitoring.
Auto alert generation in caseof emergency medical episodes
(e.g. Heart attacks, seizures).
181. CurrentFocusAreas (contd.)
Pollution and CalamityMonitoring
Monitoring for weather or man-made basedcalamities.
Alert generation in case of above-threshold pollutants in the air
or water.
Resource reallocation and rerouting of services in the event of
calamities.
18
1
182. CurrentFocusAreas (contd.)
Smart Energy
Smart metering systems.
Smart energy allocation and distribution system.
Incorporation of traditional and renewable sources of
energy in the same grid.
183. CurrentFocusAreas (contd.)
SmartAgriculture
Automatic detection of plant waterstress.
Monitoring of crop healthstatus.
Auto detection of cropinfection.
Auto application of fertilizers andpesticides.
Schedulingharvesting and arranging proper transfer of
harvests to warehousesor markets.
18
3
184. IoTChallenges inSmartCities
Security andPrivacy
Exposureto attacks (e.g. cross-sitescripting, side channel,etc.).
Exposureto vulnerabilities.
Multi-tenancy induces the risk of dataleakage.
Heterogeneity
Integration of varying hardware platforms andspecifications.
Integration of different radiospecifications.
Integration of various softwareplatforms.
Accommodating varying user requirements.
185. IoTChallenges inSmartCities (contd.)
Reliability
Unreliable communication due to vehicle mobility.
Devicefailures still significant
Largescale
Delaydue to large scaledeployments.
Delaydue to mobility of deployednodes.
Distribution of devicescanaffect monitoring tasks.
18
5
186. IoTChallenges inSmartCities (contd.)
Legaland Socialaspects
Servicesbasedon user provided information may be subject to
localor
international laws.
Individual and informed consent required for using humansasdata
sources.
18
6
187. IoTChallenges inSmartCities (contd.)
Bigdata
Transfer,storageand maintenance of huge volumes of data is
expensive.
Datacleaning and purification is time consuming.
Analytics on gigantic data volumes is processingintensive.
188. IoTChallenges inSmartCities (contd.)
SensorNetworks
Choiceof appropriate sensorsfor individual sensingtasksis crucial.
Energyplanning is crucial.
Deviceplacement and network architecture is important for
reliable
end-to-end IoTimplementation.
Communication medium and means play an important role in
seamlessfunction of IoTin smart cities.
18
8
189. SmartCities-DataFusion
Enormousvolume of datais produced periodicallyin asmart city
environment.
Challengesinclude makingthe available/incominglargedata
volume precise andaccurate.
Quality of data precisionand accuracyaffects the quality of decision
making in IoT-enabled smartcities.
Datafusion enablesoptimum utilization of massivedata gathered
from multiple sources,and acrossmultiple platforms.
Source: Alam, Furqan, et al. "Data Fusion and IoT for Smart Ubiquitous
2
190. Multi-sensorData Fusion
Combinesinformation from multiplesensorsources.
Enhancesthe ability of decisionmakingsystemsto includea
multitudeof variablesprior to arrivingat adecision.
Inferencesdrawn from multiple sensortype datais
qualitatively superior tosinglesensortype data.
Information fusion generatedfrom multiple heterogeneous sensors
providesfor betterunderstandingof the operational surroundings.
Source: Alam, Furqan, et al. "Data Fusion and IoT for Smart Ubiquitous
Environments: A Survey." IEEE Access (2017). 3
191. Challenges
Imperfection Inaccurate or uncertain WSNsensordata
Ambiguity Outliers, missingdata
Conflicts Samesensortype reports different data for the samelocation.
Alignment Arises when sensor data framesare converted to asingular frame prior to
transmission
Trivialfeatures Processingof trivial data features may bring down the accuracyof the whole
system
Source: Alam, Furqan, et al. "Data Fusion and IoT for Smart Ubiquitous Environments: A Survey." IEEE Access(2017).
4
192. DataFusionOpportunitiesin IoT
Collectivedataisrich in information andgeneratesbetter
intelligencecomparedto datafrom singlesources.
Optimal amalgamation ofdata.
Enhancingthe collective information contentobtainedfrom multiple
low-power, low-precisionsensors.
Enableshidingof critical datasourcesandsemantics(useful in military
applications, medical cases,etc.).
Source: Alam, Furqan, et al. "Data Fusion and IoT for Smart Ubiquitous
Environments: A Survey." IEEE Access (2017).
5
193. StagesofData Fusion
Decision level
• Ensemble of
decisions
Feature level
• Fusionof
information
prior to
decision
making
Pixel level
• Fusionof
information
at the
imaging
device level
itself
Signallevel
• Fusionof
information
at thesensor
node/ within
the local
network
itself.
Source: Alam, Furqan, et al. "Data Fusion and IoT for Smart Ubiquitous Environments: A Survey." IEEE Access(2017).
6
194. MathematicalMethodsofData
Fusion
Probability based
• Bayesiananalysis, Statistics, Recursive methods
AI based
• ANN,Machine Learning, CNN
Theory of Evidencebased
• Belief functions, Transferable belief models
Source: Alam, Furqan, et al. "Data Fusion and IoT for Smart Ubiquitous Environments: A Survey." IEEE Access(2017).
7
197. SmartParking
Shortensparking searchtime ofdrivers.
Reducestraffic congestion.
Reducespollution by keepingunnecessarilylingeringvehicles off the
roads.
Reducesfuel consumption andcosts.
Increasesurbanmobility.
Shorterparkingsearchtime resultsin more parked time, and hence,
more revenue.
Source: Lin, Trista, Hervé Rivano, and Frédéric LeMouël. "A Survey of Smart
Parking Solutions." IEEETransactionson Intelligent Transportation
10
206. SmartHome-Introduction
Smart home infrastructure consistsof:
Intelligent networking deviceinfrastructure
Seamlessintegration of various devicesusingwired/wireless
technologies
Allows easeof usefor householdsystems.
Createsahighly personalizedandsafehome space
Corporations seriously indulging in smart homesystems include
GE,Cisco,Google,Microsoft, andothers.
2
207. SmartHome
Providesproductive and cost-efficientenvironment.
Maximizesthe effectiveness of the occupants.
Providesefficient management with minimum life-time costs
of hardware andfacilities.
Optimizes-
Structures
Systems
Servicesand management
Interrelationships between the abovethree
3
Introduction to Internet ofThings
208. SmartHome Aspects
Source:Toschi,Guilherme Mussi, LeonardoBarreto Campos,and Carlos EduardoCugnasca."Home automation networks:Asurvey." Computer
Standards& Interfaces 50 (2017):42-54.
4
Introduction to Internet ofThings
210. HANElements
Internet Protocol(IP)
Multi-protocol gateway bridges
non-IPnetwork toIP network.
Bridging between new
technologies islimited.
For new technologies or
networks, anew mappingis
required for bridging to
perform satisfactorily.
Internet Protocol
211. HANElements
Wired HAN
Easyintegration withpre-
existing houseinfrastructure.
Lowcost.
Canusepower lines, coaxial
cables,telephone lines,optical
fibers, and other such
technologies for
communication.
Wired HAN
7
212. HANElements
Wireless HAN
Canusepopular home Wi-
Fi, ZigBee,and evennew
standards, suchas
6LoWPAN.
Wireless makes
implementation
easy.
WirelessHAN
8
214. HANStandards
Universal Plugand Play(UPnP).
Application layer technology,mainly
web-based.
TCP/IPprotocol stackprovides support
for the lower layers,and enables seamless
integration of various technologies.
Providestransparent networking with
support for zero-configuration
networking and automatic discoveryof
devices.
UPnP
DLNA
Konnex
LonWorks
Zigbee
X-10
10
215. HANStandards
Digital LivingNetworkAlliance
(DLNA)
Tradeorganization created bySony,
Intel, andMicrosoft.
Connectscable-basednetworks with
wireless networks for increased
sharing of media,control and access.
Domestically sharesnetwork media
resources.
UPnP
DLNA
Konnex
LonWorks
Zigbee
X-10
11
216. HANStandards
Konnex(KNX):anopen
important standard for home
and building networks.
Utilizes the full rangeof home
communication infrastructure –
Power lines, coaxialcables,twisted
pair, RF
,etc.
Must be setup and configured viaa
software before its properusage.
UPnP
DLNA
Konnex
LonWorks
Zigbee
X-10
217. HANStandards
LocalOperation Networks(LonWorks).
Everydevice includes aNeuron Chip,a
transceiver and the application electronics.
Neuron chip is aSOCwith multiple
microprocessors,RAM,ROMand IO
interface ports.
Splits device groups into intelligent
elements, which can communicate through
aphysicalcommunication medium.
UPnP
DLNA
Konnex
LonWorks
Zigbee
X-10
13
218. HANStandards
Zigbeeconsistsof four layers–Physical,
MediumAccessControl, Network, and
Application.
Physicaland MAClayers are defined by
IEEE802.15.4,whereas Network and
Application are defined byZigbee.
Aims at low-cost, low-energydevices.
ZigBeeAlliance is composed of
Mitsubishi, Honeywell,Invensys,
Motorola andPhilips
UPnP
DLNA
Konnex
LonWorks
Zigbee
X-10
14
219. HANStandards
X-10enables remote control of compliant
transmitters and receivers over power lines and
electrical wirings present in the house.
Adopted by GEandPhilips.
Standarddefines procedures for
transmission of bits overACcarrier
signals.
Low-speed and low data rate.
Mainly usedfor control of lighting, appliance
networks and security sensors.
UPnP
DLNA
Konnex
LonWorks
Zigbee
X-10
15
220. HANArchitectures
UsesXMLfor description and web- services
for control.
Follows aServiceorientedArchitecture
(SOA).
Not tied toanysoftware, languageor
architecture.
Acentral gateway connects different
technologies.
Atech Manager for eachtechnology provides
web servicesfor controland access.
DomoNet
Jini
16
221. HANArchitectures
Connects various devices sharing their resources
with auto-configuration and auto-installation.
Basedon JA
V
Aenvironment andpursued
by SunMicrosystems (Now, Oracle).
Constructs an organized distribution system
without acentral node (federation).
Jini appsusebytecode to run JVM,and are
portable.
Follows Object Oriented Paradigm.
DomoNet
Jini
17
222. HANInitiatives
Middleware for embedded
intelligent systems.
ConnectsaService Oriented
Architecture Network.
Connecteddevicesmay havelimited
resources,low processing power,
memory or energyconsumption.
Eachdevice hasan embedded HYDRA
client which acts asaproxy between
the device and themiddleware.
Project
HYDRA
Amigo
18
224. HANInitiatives
Aimed at:-
Ambient intelligent systems
For networked homesystems
Features user-friendly interfaces,
interoperability, and automatic
discovery of devices and
services.
Project
HYDRA
Amigo
20
227. Introduction
Vehiclesequipped with
Sensors
Networking and communicatingdevices
Capableof :
Communicating with other deviceswithin the vehicle
Communicating with other similarvehicles
Communicating with fixedinfrastructure
Source:Kim,Younsun,HyunggoyOh,and SunghoKang."Proof of Conceptof HomeIoTConnectedVehicles." Sensors17.6 (2017): 1289.
2
228. Challenges
Security
Privacy
Scalability
Reliability
Quality of service
Lackof globalstandards
Source:Kim,Younsun,HyunggoyOh,and SunghoKang."Proof of Conceptof HomeIoTConnectedVehicles." Sensors17.6 (2017): 1289.
3
230. Vehicle-to-Everything(V2X)Paradigm
Main component of futureIntelligent
Transportation System(ITS).
Enablesvehiclesto wirelessly sharea
diverserangeof information.
Information sharing maybe with other
vehicles, pedestrians, or fixed
infrastructures (mobile towers,parking
meters, etc.)
Allows for traffic management,
ensuringon-road andoff-roadsafety,
mobility fortraveling.
5
231. V2X
Follows adistributed architecture, where contents are widely
distributed over the network.
Not restricted to single source informationprovider.
Designedmainly for highly mobileenvironments.
Canshareinformation to nodes in vicinity, aswell asremotelylocated.
Hasgreatly enhanced travel efficiency, aswell assafety.
Thenetwork is mainly usedasatool for sharing and
disseminating information.
Source:Zhu, Z.,et al. "Recent advancesin connected vehicles via information-centric networking." Intelligent and ConnectedVehicles (ICV2016),
IETInternational Conferenceon. IET
,2016.
6
Introduction to Internet ofThings
232. FailuresofTCP/IPinV2X
Designedmainly for handling information exchangebetween a
singlepair of entities.
Information exchangedependent on the location ofdata.
Canonly identify the addressesof endpoints, which aloneis not
useful for contentdistribution.
Increasein number of wirelessdevices,restricts themobility of
thenodes.
Source:Zhu, Z.,et al. "Recent advancesin connected vehicles via information-centric networking." Intelligent and ConnectedVehicles (ICV2016),
IETInternational Conferenceon. IET
,2016.
7
233. ContentCentricNetworking(CCN)
CCNis derived from Information Centric Networking (ICN)
architecture.
Focusesmore on the data than its actuallocation.
Hierarchically named data.
Hierarchical data is transmitted directly instead of being
part ofa conversation.
Enablesscalable and efficient datadissemination.
In-network caching allows for lowdata traffic.
Works well in highly mobileenvironments.
Source:Zhu, Z.,et al. "Recent advancesin connected vehicles via information-centric networking." Intelligent and ConnectedVehicles (ICV2016),
8
234. VehicularAd-hoc Networks(VANETs)
Basedon:
Dedicated Short-Range
Communication (DSRC)
WirelessAccessin
Vehicular Environment
(WAVE)
Routing protocols derivedfrom
MANETs.
High throughput achievable inmobile
environments.
Guaranteed low-latency in mobile
environments. 9
235. VANETFeatures
HighDynamic Topology
• Vehiclesin highly mobile environments causesconstant changesin networkpartitioning
and topology.
Hightransmissionandcomputation capability
• Vehicle-stored energy sourcesand computational power do notrestrict capabilities.
Unstableconnectivity
• Link durations are short due to highly dynamic nature of VANET
s.
LargescaleCanbe easily scaledup to include all vehicles onroads.
Predictablemobility pattern
• Vehicular restriction within roads, makesmobility patternpredictable.
Source:Zhu, Z.,et al. "Recent advancesin connected vehicles via information-centric networking." Intelligent and ConnectedVehicles(ICV2016),
IETInternational Conferenceon. IET
,2016.
10
236. ApplicationsofVANETs
Safety
• Emergencybraking, lane change warning, collision avoidance,hazard
notification
Efficiency
• Congestion management, electronic toll collection, parkingavailability
Commercial
• Internet access,multimedia stream
Comfort
• Weather information, autonomous driving, journey timeestimation
Source:Zhu, Z.,et al. "Recent advancesin connected vehicles via information-centric networking." Intelligent and ConnectedVehicles (ICV2016),
IETInternational Conferenceon. IET
,2016.
11
Introduction to Internet ofThings
237. CCN forVANETs
Routing
Forwardingandrouting basedon nameof content (not location).
Individual content’s nameprefixesare advertised by routers acrossthe
network.
Thishelpsto build aForwardingInformation Base(FIB)for eachrouter.
Thenameof content remains sameand uniqueglobally.
Noissuesof IPaddressmanagement or addressexhaustion.
Communication doesnot dependon speedor direction of nodes.
12
238. CCN forVANETs
Scalability
Anin-network caching mechanism at eachrouter.
Uniquely identifiable (named) data chunksare stored in ContentStore
(CS),which acts asa cache.
Subsequentrequestsfor astored data chunk canbe made to aCS.
Thenaming systemin the CSenables adata to be usedmultipletimes,
unlike normal IP-basedrouters.
Reducednetwork load during increasednetwork size,asaresult of the
caching mechanism.
240. Bodyand BrainArchitecture
An in-vehicle networking architecture.
Three layered architecture.
Thebody consists of intelligent networking nodes
(INN) which constantly collect information from
the vehicle.
Thebrain manages central
• coordination.
246. TechnologicalBackground
TheUSDepartment of Transportand FederalCommunications
Commission allocated 75MHz(5850-5925MHz) asthededicated
spectrum for ICVs.
It isbasedon DedicatedShort RangeCommunication(DSRC)
technology.
IEEEdeveloped IEEE802.11p and IEEE1609 as DSRCstandards.
Societyof Automotive Engineers(SAE)cameup with SAEJ2735and
J2945asDSRCstandards.
Source:Li,Yan,et al. "Big wave of the intelligent connected vehicles." ChinaCommunications 13.Supplement2 (2016): 27-41.
3
247. IEEE1609Family
IEEEP1609.0Draft Standardfor WirelessAccessin Vehicular Environments(W
A
VE)-
Architecture
IEEE1609.1-2006- TrialUseStandard for WirelessAccessin Vehicular Environments
(W
A
VE)- ResourceManager
IEEE1609.2 -2006- Trial UseStandard for WirelessAccessin VehicularEnvironments
(W
A
VE)- Security Servicesfor Applications and ManagementMessages
IEEE1609.3 -2007 - Trial UseStandard for WirelessAccessin VehicularEnvironments
(W
A
VE)- Networking Services
IEEE1609.4 -2006- Trial UseStandard for WirelessAccessin VehicularEnvironments
(W
A
VE)- Multi-Channel Operations
IEEEP1609.11Over-the-Air DataExchangeProtocol for Intelligent Transportation
Systems(ITS).
4
257. InfrastructureDomain
RSUsconnected to Internet by meansofGateways.
In the presence of RSUs, the vehicles may communicate to the
Internet via V2Iinterfaces.
In the absence of RSUs,the vehicles may communicate with each
other or the Internet through cellular networks such as 3G/4G,
L
TE,etc.
14
264. “IoTas a concept has crossed the chasm from slideware
to reality with many industries implementing IoT
solutions.”
‐ Paul Howarth, Senior Manager, Corporate Development,CISCO
2
265. Introduction
3
The main aim of Internet ofThings (IoT) is
to globally connect smart ‘things’or‘objects’.
objects are uniquelyidentified.
interoperability among theobjects.
The Industrial Internet of Things (IIoT) is an application of IoT in
industries to modify the various existing industrial systems. IIoT
links the automation ,,planning and productlifecycle.
268. Introduction (contd.)
IIoT includes–
machine learning
big datatechnology
machine ‐ to ‐ machine interaction(M‐2‐M)
automation.
IIoT is supported by huge amount of data collected from
sensors. It is based on “wrap & re‐use” approach, rather than
“rip& replace”approach.
(Source : http://www.mhi.org)
269. Introduction (contd.)
Industrialization
(1870)
Power Generation
& Mechanical
Automation (1782)
Electronic
Automation
(1969)
Smart
Automation
(today)
1st Industrial Revolution :Mechanized
production
2nd Industrial Revolution :Mass
production
3rd Industrial Revolution :Internet
evolution and automation
4th Industrial Revolution :IIoT
Fig 2: Industry 4.0
Source:http://www.industry40wood.com
271. Introduction (contd.)
IIoT is a networkof
physicalobjects
systems
platforms
applications
These networks can communicate with each other,externalenvironment
and otherpeople.
The acquisition of IIoT has led to availability and affordability of sensors,
processors, and other technologies which facilitates capture and access to
real‐time information
274. Design Considerations
Touse an IoTdevicefor industrial applications,the followingdesign
objectivesare to be considered–
Energy:TimeforwhichtheIoTdevicecanoperatewithlimitedpower
supply.
Latency:Timerequired to transmit thedata.
Throughput :Maximum datatransmitted across thenetwork.
Scalability :Number o
fd
e
e
v
i
c
e
ssupported.
Topology:Communication among the devices,i.e.interoperability.
SafetyandSecurity:Degreeofsafetyandsecurityoftheapplication.
275. IoT
• Focused on convenience
of individuals
• M‐2‐M communication:
Limited
• Applications areasare
at consumer‐level
IIoT
• Focused on efficiency,
safety and security of
the operation.
• M‐2‐M communication:
Extensively.
• Application areas are at
industries.
Difference between IoT and IIoT
The main differences between IoT and IIoT are :
276. Difference between IoT and IIoT
(contd.)
Devices
Network
(connectivity)
Service
enablement
Application and
data
System
integration
M‐2‐M focus
IoTfocus
277. Service Management in IIoT
“Service management refers to the implementation and
managementof the quality of services which meets the
end‐users demand”
“Serviceis a collection of data and associated behaviors to
accomplish a particular functionor featureof a device or
portions of a device”.
278. Service Management in IIoT
Service canbe of two types,which are ‐
Primary service ‐The basic services which are
responsible for the primary node functions are termed
as primary service.
Secondary service ‐The auxiliary functions which
provide servicesto the primary service or secondary
services are termed as secondary service.
279. Applications ofIIoT
The keyapplication areas of
IIoTare ‐
Manufacturing industry
Healthcare Serviceindustry
Transportation & logistics
Mining
Firefighting
280. Manufacturing Industry
3
The devices, equipment, workforce, supply chain, work
platform are integrated and connected to achieve smart
production. This will led to –
reduction in operationalcosts
improvement in the productivity of theworker
reduction in the injuries at theworkplace
resource optimization and wastereduction
end‐to‐end automation.
281. Healthc are ServiceIndustry
Patients can be continuously monitored due to the
implanted on‐body sensors. This has led to–
improved treatmentoutcome
costs hasreduced
improved diseasedetection
improved accuracy in the collection ofdata
improved drugs management.
282. Transportation & logistics
Toimprove safety,efficiency of transportation,Intelligent
Transportation system (ITS) is developed which consists of
connected vehicles. ITS provides –
Vehicle – to – sensor connectivity
Vehicle – to – vehicle connectivity
Vehicle – to – internet connectivity
Vehicle – to – road infrastructure
Dedicated short‐range communications (DSRC) is the key
enabling technologyforV2V andV2R communications.
283. Transportation & logistics
In IIoTscenario thephysical objectsare providedwith
bar codes
RFID tags
hence, real‐time monitoring of the status and location of the
physical objects fromdestination to the origin, across the
supply chain is possible.
Security and privacy of the datashould be maintained.
284. Mining
Toprevent accidents inside the mines ‐RFID,Wi‐Fi and
other wireless technologies are used,which
provides early warning of anydisaster
monitors air‐quality
detects the presence of poisonous gases inside the
mines
oxygen level inside themines.
285. Firefighting
Sensor networks, RFID tags are used to perform
automaticdiagnosis
early warning ofdisaster
emergencyrescue
provides real‐timemonitoring Hence, improves public
security.
286. Examples of IIoT
Introduction to Internet of Things
9
Examples of IIoT are‐
unmanned aerial vehicles (UAVs) to inspect oilpipelines.
monitoring food safety usingsensors.
minimizing workers’exposure to noise, chemicals and
otherhazardous gases.
unmanned marine vehiclewhich can collectdataup to a
year without fuel or crew.
287. Connected Ecosystems in IIoT
scenario
Traditional supplychains in industries are linear in nature.
Toshift the business focus from products to outcomes, new
ecosystem should befollowed.
Digital ecosystems progress ata much faster ratethan
physical industries. Hence, itcan quickly adapt to the changes
in the external environments.
288. Integration of Digital and
Human Workforce
In IIoT,machines become more intelligent.Hence, the
automatedtasks canbe done in the industries atlower costs
and higher quality level.
Humans will work with machines, the outcome will be
higher overall productivity.
IIoTwill reformand redefine the skills of the workers.
289. Creation of New Jobs
The creation of new composite industries, such as precision
agriculture, digital healthcare system, digital mines etc., will
lead to developmentof new jobopportunities.
Highly automated machines will require lesser number of
unskilled workers, but will require skilled experts with
digitaland analyticalskills.
290. Reformation of Robots
In IIoTenvironment, robots are featured with three capabilities :sensing,
thinking and acting.They will be reformed with the ability to carry out
repetitivetasks.
Robots will be more intelligent butwill work under the supervision of
human beings. Their availability willincrease.
Robots will be reprogrammable toperform nasekws.t They havethe
capability to ‘learn’faster.
291. Challenges in IIoT
Primary challenges
Identification of objectsor
things
Managehuge amount of data
Integrate existing
infrastructures into newIIoT
infrastructure
Enabling datastorage
292. Challenges in IIoT(contd.)
SafetyChallenges
Worker health andsafety
Regulatory compliance
Environmental protection
Optimized operations
293. Challenges in IIoT(contd.)
Hazards (related)
Handling, storing or using hazardoussubstances
Oxygendeficiency
Particulates
Radiation
Physiological stress
294. Challenges in IIoT(contd.)
Standardization
Standardization plays an important role in the development of
the system.
Goal:Toimprove the interoperability of the different systems/
applications and allow the products/services to performbetter.
295. Challenges in IIoT(contd.)
Standardization
The problems related to
standardizationare:
Interoperability
Semantic interoperability (data
sematics)
Security andprivacy
Radio access levelissues.
296. Challenges in IIoT(contd.)
Privacy and securityissues
The two most important concerns relatedwithIIoTare ‐
information security
data privacyprotection
The devices/thingscan be tracked,monitored and connected.
So there are chances of attackon the personal and privatedata.
297. Challenges in IIoT(contd.)
Privacy and securityissues
Examples –
Healthcare industry – the medical data of a patient
must not be tampered, or altered by any person in the
middle.
Food industry – the deterioration of any food item being
sent to the company must be kept confidential as it will
affectthe reputation of the company.
298. Risksassociated with IIoT in
Manufacturing
Though IIoTprovides new opportunities, but few
factors may cause hindrance in the path tosuccess,
which are :
lackof vision andleadership
lackof understanding of values among management
employees
costly sensors
inadequate infrastructure.
299. Meet the challenges: Sensor
improvement
Improvement in sensor technologies–
miniaturization
performance
cost and energyconsumption.
300. Meet the challenges : Manufacturing
Manufacturers use software capabilities to improve
operationalefficiency through –
predictivemaintenance
savings on scheduledrepairs
reduced maintenancecosts
reduced number ofbreakdowns.
301. Case study : RtTech Software
24
Rt Techparticularizes in software which–
improves industrial facilities’efficiency
improves productivity.
Energymanagement solution,which leads to reductionin
the plant’s highest variablecost.
RtTechautomates the process of mapping and
managing energy consumption.
302. PRODUCTS DEVELOPED
M‐2‐M communication :Intelligent Radio Modem(IRM)
IRM 1500 &ACE 1000 ‐IRM
simple
M‐2‐M connectivity
datatransmission
These devices provide easy maintenance and installation.They
canbe connected to IP and non‐IP serial devices to extend the
capability to monitor and communicate with othertechnologies.
303. PRODUCTSDEVELOPED (contd.)
Comtrol – IO Link Master Gateway
It can be easily integrated into the
industrial network with existing and
newinstallations.
It supports Ethernet/IP,
PROFINET(PNIO) and
ModbusTCP
.
304. Benefits of IIoT
The benefitsof IIoTare
Improved
connectivity
among devices
Improved
efficiency
Upgraded
scalability
Reduces operation
time
Remote diagnosis
Costeffective
27
305. Recent Research trends in IIoT
Recent research challenges in IIoT are‐
Toimprove the communications among the different things or
objects.
Todevelop energy‐efficient techniques so as toreduce power
consumption bysensors.
Todevelop context‐aware IoT middleware for better understanding
of the sensor data.
Tocreatesmart objects with larger memory,
processingaasnodnirneg capabilities.
306. Conclusion
IIoT system requires the following:
Smaller,less expensive sensors which makes them
easilyaccessible.
Distributed control of assembly line,automated
monitoring,control and maintenance.
308. Future of IoTapplication inagriculture
Image template source: https://pixabay.com/p‐747175/?no_redirect
Soil moisture and water level monitoring
Automated irrigation system
Automation in Recycling of Organic Waste and
Vermicomposting
Automated sowing and weeding system
2
309. Future of IoT application inagriculture
Soil moisture and water level monitoring.
Automated irrigation system.
Automation in Recycling of Organic Waste and
Vermicomposting.
Automated sowing and weeding system.
311. AgriSens: Smart Water Management using IoT
4
Objectives
More yields with lesswater
Save limited water resource in acountry
Automatic irrigation
Dynamic irrigation treatments in the different phases of a
crop’slife cycle
Remote monitoring andcontrolling
Source: Project name: Development of a Sensor based Networking System for Improved Water Management for Irrigated Crops, funded by MHRD, Govt. of India
312. AgriSens: Smart Water
Management using IoT (Contd.)
Proposedarchitecture
Sensing and actuatinglayer.
Processing, storage, andservice layer.
Application layer.
5
313. AgriSens: Smart Water
Management using IoT (Contd.)
6
Design
Integrated design forsensors
Integrated design for sensornode
Integrated design for remoteserver
314. AgriSens: Smart Water
Management using IoT (Contd.)
Integrated design forsensors
Fig 4: Designed water‐levelsensor
Fig 5: EC‐05 soil moisturesensor
Source: Project name: Development of a Sensor based Networking System for Improved Water Management for Irrigated Crops, funded by MHRD, Govt. of India
7
315. AgriSens: Smart Water
Management using IoT (Contd.)
Integrated design for sensornode
Fig 2:The blockdiagram of a sensor node
8
316. AgriSens: Smart Water
Management using IoT (Contd.)
Integrated design for sensornode
Fig 3: Designed sensornode
Source: Project name: Development of a Sensor based Networking System for Improved Water Management for Irrigated Crops, funded by MHRD, Govt. of India
9
317. AgriSens: Smart Water
Management using IoT (Contd.)
10
Integrateddesignforremoteserver
Repositorydataserver:Communicates with the deployed IoTgateway
in the fieldby using GPRStechnology
Web server:Toaccess field dataremotely
Multi users server:Sends field information to farmer’s cellusing SMS
technology and also executes farmer’s query andcontrollingmessages
318. AgriSens: Smart Water
Management using IoT (Contd.)
11
Implementation
Field demo
Website demo
Project details fromwebsite
319. AgriSens: Smart Water
Management using IoT (Contd.)
Results
Fig. 6:Average soilmoisture
12
Source: Project name: Development of a Sensor based Networking System for Improved Water Management for Irrigated Crops, funded by MHRD, Govt. of India
Vegetativephase Reproductive phase Maturityphase
Editor's Notes
Source: Zhu, Z., et al. "Recent advances in connected vehicles via information-centric networking." Intelligent and Connected Vehicles (ICV 2016), IET International Conference on. IET, 2016.
Source: Zhu, Z., et al. "Recent advances in connected vehicles via information-centric networking." Intelligent and Connected Vehicles (ICV 2016), IET International Conference on. IET, 2016.