IOT and Applications

GTU # 3160716

Unit-4
Cloud for IoT

Prof. Kalpesh H Surati

Computer Engineering Department
Darshan Institute of Engineering & Technology, Rajkot
[email protected]
+91 99250 10033
Introduction to Cloud Computing
 Cloud computing is the delivery of computing services
over the internet, by providing flexible, affordable,
effective and efficient resources for development.
 That includes servers, storage, databases, networking,
software, analytics, and intelligence.
 Cloud computing provide economical freedom along
with technical strength to the application.
 In other word, it is about outsourcing of IT services and
infrastructure to make them available anywhere via the
Internet.

Application logic runs Application logic runs

on user’s computer in the cloud
Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 2
Service Models of Cloud Computing
 There are three main service models
of cloud computing – Infrastructure as
a Service (IaaS), Platform as a Service
(PaaS) and Software as a Service
(SaaS).
 Each model represents a different part
of the cloud computing stack.
 Each type of cloud service provides
you with different levels of control,
flexibility, and management.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 3

Example of “as a Service”
Pizza as a Service
Traditional Infrastructure Platform Software
On-Premises as a Service (IaaS) as a Service (PaaS) as a Service (SaaS)

You Manage

Vendor Manage

Made at Home Take and Back Pizza Dined at

Delivered Restaurant

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 4

Example of “as a Service”

Vehicle
as a
Service

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 5

Cloud Services

Traditional Infrastructure Platform Software

On-Premises as a Service (IaaS) as a Service (PaaS) as a Service (SaaS)

Applications Applications Applications Applications

Data Data Data Data

Runtime & Middleware Runtime & Middleware Runtime & Middleware Runtime & Middleware You Manage

OS OS OS OS
Vendor Manage
Virtualization Virtualization Virtualization Virtualization

Networking Networking Networking Networking

Storage Storage Storage Storage

Servers Servers Servers Servers

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 6

Software-as-a-Service (SaaS)
 Complete software application as a
service is provided to the user that is
run and managed by the service
provider.
 It can also be called application as a
service as a pay monthly, yearly etc.
subscription.
 In SaaS, user don't need to worry
about software upgradation and
management.
 A common example of a SaaS
application is web-based email
where you can send and receive
email without having to manage the
server that the email program is
running.
Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 7
Platform-as-a-Service (PaaS)
 It provides a development environment to
application developers.
 Operating system, programming-language
execution environment, database, web server,
development tools, APIs, libraries, etc., will be
provided by the cloud service provider.
 Users have to build, manage and maintain the
applications.
 Application developers develop and run their
software on a cloud platform instead of
directly buying and managing the essential
hardware and software layers.
 AWS Lambda, Google App Engine, IBM Cloud
Foundry, Oracle Cloud Platform, Red Hat
OpenShift, Zoho Creator.
Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 8
Infrastructure-as-a-Service (IaaS)
 IaaS provides access to computing
hardware, networking features, and data
storage space.
 Where the consumer is able to deploy
and run software, which can include
operating systems and applications.
 The consumer does not manage or
control the underlying cloud
infrastructure but has control over
operating systems, storage, and
deployed applications
 This service provides you with the
highest level of flexibility.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 9

Deployment Models of Cloud Computing
 Deployment models of cloud
computing are categorized based
on their location as public, private
or hybrid cloud
 It is often possible to choose a
geographic area to put the data
“closer” to users.
 More the money you pay, better in
comfort and service

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 10

Private Cloud
 Private cloud provides computing services within the
organization’s private network and selected other users.
 In this cloud model all the hardware, software, datacenter,
employees, infrastructure, etc., are maintained, monitored,
and installed by the organization.
 This particular deployment model can be chosen wherever
confidentiality matters the most.
 Advantages
 High level of security and privacy
 More control flexible in terms of deciding and managing the
resources
 Disadvantages
 Very Expensive
 Need technical skill for maintaining and difficult to management
 Policies and other related things are to be framed carefully to
make sure that the data is safe
Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 11
Public Cloud
 The cloud resources are owned and managed by
a third-party cloud service provider.
 Pay as per usage approach makes it most cost
effective model.
 Advantages
 Inexpensive model, no need to invest in setting up
the infrastructure and maintenance
 Less technical skill required
 Customer support team can be reached on demand
 Easily scaled up or scaled down based on
requirements
 High reliability—a vast network of servers ensures
against failure.
 Disadvantages
 Security and privacy issues are the major challenges
 Less flexibility and controls
Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 12
Hybrid Cloud
 This deployment is a mix of both private and
public cloud deployment
 The resources offered and managed are both in-
house and third party based
 Advantages
 Less investment needed to setup the infrastructure
 Less technical skill required to manage and maintain
the cloud
 Customer support team can be reached on demand
 Easily scaled up or scaled down based on
requirements
 Disadvantages
 Security and privacy issues are the major challenges
 Less flexibility and controls compare to public cloud

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 13

IoT with Cloud - Challenges
 We have already discussed the
Privacy and
Security
challenges of IoT and cloud.
 These challenges can be increased
Expertise
Bandwidth when the IoT and cloud are
Cost
integrated.
 Here we will discuss seven such
challenges
IoT Cloud  Privacy and Security
Challenges Migration  Bandwidth Cost
Data
and
Ownership
Portability  Migration and Portability
 Availability, Reliability and Robustness
 Costing
Availability,
 Data Ownership
Costing
Reliability
and
 Expertise
Robustness

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 14

Privacy and Security
 Security is a major concern in the field of IoT.
 Valuable data goes into the cloud, outside the firewall this data becomes hackable.

The following are solutions to this privacy and

security challenge.
 Periodic monitoring of the network activities
 Select private cloud if the data is confidential
 To reduce the risk of being exposed, use
recognized antivirus solutions.
 Before signing the contract with a cloud service
provider, it is necessary to read and understand
the regulations involved in the service being
provided.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 15

Bandwidth Cost
 Bandwidth is one of the challenges
because of continues data transferring
from the IoT devices.
 IoT is all about data, and in most cases,
this would be big data. Hence, huge
investment in storage is needed.
 Cloud computing is preferred for
storage and processing in IoT.
 Small-scale IoT application demanding
lesser resources.
 But if the application is data
concentrated, then the investment in
bandwidth would be considerable.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 16

Migration and Portability
 When data is to be moved to or migrate
from the cloud, we have to take care of the
followings.
 How easy and safe is it to move the data?
 How much downtime would this process
require?
 What is the strategy to migrate data to the
cloud?
 Will it be easy to select out of the cloud and
take data back to the infrastructure?
 How much would it cost?
 Would there be support offered to migrate
smoothly to another cloud service provider?
 All these challenges are doubled with IoT as
the data comes from the various sensory
nodes at a very high speed.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 17

Availability, Reliability and Robustness
 Continuous monitoring and
reading of the data need to have
continuous cloud service
availability in IoT.
 In downtime, it would miss critical
data so reliability of the process
has to be monitored.
 The process should be robust
towards handling data at different
rates.
 Data could be flooded or slowed at
anytime, in the both situations it
should be handled effortlessly.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 18

Costing
 One of the main advantages of cloud is that it can scale up with rising demand.
 While it is scalable and flexible, an organization should plan its budget carefully.
 Wrong selection for subscription without having clear vision and planning, it may lead to
unnecessary cost.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 19

Data Ownership
 The data stored by the user on the cloud is
owned by the user.
 This means that the data is under the
ownership of the person who generates it.
 However, when opting for cloud storage, the
data is under the custody of the cloud service
provider.
 Then, it appears that the service provider
owns the data.
 When it comes to IoT, the data is generated at
multiple points and ownership could lie with
multiple participating parties.
 Hence, in IoT the ownership-related
challenges are multiplied.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 20

Expertise
 To use the cloud with IoT requires a
specific skill set.
 The cloud platform gets updated
every now and then and so experts
have to constantly upgrade
themselves.
 Expertise is a definite challenge.
 When IoT and cloud comes together
then it will be more challenging,
 To understand the sensors and at
the same time, to be updated on
cloud development is a challenging
task.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 21

Selection of Cloud Service Provider: An Overview
 There are many parameters to select cloud service providers from the numerous service
providers.
 It is advisable to consider following parameters while selecting the cloud service provider.
 Certification and Standards Compliance
 Financial Health of the Service Provider
 Business and Technology Strength
 Compliance Audit
 Service Level Agreements
 Reporting/Tracking
 Costing and Billing
 Maintenance Monitoring and Upgrade
 Support
 Security

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 22

Criteria 1: Certification and Standards Compliance
 When a product adheres to the standards that
are accepted widely, it is considered as a
reliable product. Similarly,
 Cloud service providers (CSPs) are expected
to comply with standards. Because product
with standards are accepted widely, it is
considered as a reliable product
 Industry accepted standards is the first
criteria to select the CSP.
 Though there are many standards framed and
followed by the industry, some of the main
standards for cloud are ISO, Open Cloud
Consortium (OCC), IEEE, SNIA (Cloud Storage
Initiative).

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 23

Criteria 2: Financial Health of the Service Provider
 The service provider should hold sufficient funding to operate
business for a long period.
 If the service provider has healthy financial status and history
of sustenance, then it is most unlikely to shut down.

Criteria 3: Business and Technology Strength

 Having the technical expertise to sustain and adapt to a client's
requirements is a key factor in selecting a CSP.
 Having just the technical skill and strength does not help; the
CSP needs business skills as well to sustain.
 Business skills include growth planning, financial planning, and
other factors that are required to sustain in the market.
 Sustenance = Technology + Business Skills

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 24

Criteria 4: Compliance Audit
 The CSP must validate compliance with the client's requirement.
 Which should be done through a proper third party audit.
 This will enable transparency and perfect validation.

Criteria 5: Service Level Agreements

 Service Level Agreements (SLAs) provides detail
information about the services being provided.
 SLAs also indicates the real value that a customer
gets out of them.
 SLAs serve as a legal contract, define the terms &
conditions and the relationship between the two
parties.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 25

Criteria 6: Reporting/Tracking
 The service provider should be capable for issuing a complete
performance report, which also highlights the problems.
 This will enable the customer to understand the complete situation.

Criteria 7: Costing and Billing

 The costing and billing should be transparent and should
provide the complete details of the usage.
 It is expected to be automated with details of the complete
resource utilization.
 It should with having clarity along with the breakup.
 This means the billing should be transparent and for the usage
only.
 This is a major factor in selecting the CSP.
Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 26
Criteria 8: Maintenance Monitoring and Upgrade

 It should be easy and less expensive to

migrate to the CSP's environment.
 When there is an upgrade, it should be
done with ease.
 Any maintenance should be easy and
affordable.
 In short, it should be easier to install,
manage, maintain, and upgrade.
 This upgrade includes migration from
private to public to hybrid cloud, if
needed.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 27

Criteria 9: Support
 Help and assistance should be
provided when required.
 Support should be available based
on the agreements and a
dedicated resource.
 Support levels based on
complexity of problem.
 Onsite support may be needed
when clarifications cannot be
offered over phone or online.
 Thus, support is a major deciding
factor for selection of a service
provider.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 28

Criteria 10: Security

 The infrastructure, both hardware and software should be secured.

 There should be defined policies about the security that should also be shared with a
customer.
 This includes everything, from access restrictions to customer data.
 The data should be safe in case of a breakdown/failure.
 The recovery and backup options should be sound.
 The physical infrastructure has to be safeguarded as well.
 All these factors would require audit, which should be carried out by a third party.
 Security is the prime concern and cannot be ignored.
 Evaluation should start from this point.
Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 29
Introduction to Fog Computing
 IoT is all about the data. The factors that Cloud Service
affect data are the four Vs - variety, velocity,
Cloud
veracity and volume.
 All IoT applications require instant analysis Fog
and action. Computing
 Most of the time, the action would be Fog Fog Fog
corrective in nature, it would be business
critical.
 In case the data volume is high and it
reaches the cloud after some delay.
 So we may lost the opportunity to use the
data appropriately.
Edge Devices
 In such cases, fog computing serves the
solution.
Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 30
Introduction to Fog Computing
 The most sensitive data should be
analyzed in the area closer to the place
where it is generated.
 With fog computing, it is possible.
 Using fog computing we can process
the data locally and to avoid the trouble
by not sending the data to the cloud.
 Respond much faster because of data
is moving locally so data travel is
reduced considerably.
 It thus process the data in
milliseconds.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 31

Introduction to Fog Computing
 Only the required data will be sent to the cloud.
 This will be based on storage requirements and guidelines.
 Predictive analytics can also be carried out with the data stored in the
cloud.
 The fog is below cloud, which means it is closer to the elements that
generate data.
 After analysis, the data stored is pushed on to the cloud.
 Results in increased efficiency and safety both physical and asset safety.
 Some examples where faster response time is extremely important are
factory or manufacturing line, oil and gas tube lines fault analysis, on-
flight diagnosis, and healthcare.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 32

Working of Fog Computing
 Sensors/devices generate data transmit it to the middle
layer, which is very close the data source.
 These nodes in the middle layer are capable of handling
the data.
 This requires minimum power and lesser resources.
 All the data need not go to the cloud at the instant.
 Also, sensitive data gets processed very fast, which
results in an instant response.
 Fog is not meant for hefty storage. It is still the cloud
that does the task of storing big data.
 Fog is just an intermediary layer for faster data
processing, and the faster response time.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 33

Summarize the concepts
Concept of fog nodes Concept of cloud computing platform
 It receives the data feed from the  It receives the data summary from
sensors, in real-time. the fog.
 Response time is minimal, ideally in  Data prediction, data analytics, data
milliseconds. storage, etc. takes place here.
 Fog computing is transit, where data is
stored for a limited time only.
 Data is then sent to cloud as a
summary.
 It is important to note that not all data
goes to the cloud.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 34

Benefits of fog computing model
 Minimal amount of data sent to the cloud.
 Reduced bandwidth consumption.
 Reduced data latency.
 Improved data security. When limited data goes to cloud, it is easier to protect it.
 Immediate processing of data in real time (this is very much needed in industrial
applications).

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 35

Difference between Edge and Fog Computing
 Both fog and edge are concerned with the computing capabilities to be executed locally,
before passing it to the cloud.
 Both aim at reducing complete dependency on the cloud to perform computation.
 Analyzing data and processing it at the cloud is to be avoided.
 Both these reduce the time delay for making faster decision for real-time applications.
 The main difference between edge and fog computing is where data processing takes place.
 Edge computing is the computing carried out at the device itself, where all the sensors are-
connected.
 In fog computing, data processing is moved to the processors that are connected to the local
area network (LAN), making it a little farther from the sensors and actuators.
 Thus, the main difference between edge and fog computing is the distance.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 36

Cloud Computing: Security Aspects
 The security of any computing platform including cloud computing depends on
 Software security,
 Infrastructure security,
 Storage security, and
 Network security,
 If any of these is compromised, it would result in security violation and could cause
damages.
 Let us discuss these security aspects briefly.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 37

Cloud Computing: Security Aspects
1. Software Security:
 Software is the core component and plays a vital role in
presenting and ensuring a secure environment.
 If there are defects created/generated during the
development phase, it is a software security threat.
 Defects such as simple software implementation
defects, memory allocation, design issues, and
exception handling all contribute to security issues.
 This can be ensured by complete and comprehensive
testing carried out at all-stages.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 38

Cloud Computing: Security Aspects
2. Infrastructure Security:
 Making sure that the infrastructure provided by
the CSP is safe is a must.
 Third party could also contribute to the
infrastructure.
 It is extremely important to check the security
vulnerabilities with the infrastructure.
 All infrastructure related guidelines should be
mentioned clearly in the agreements and should
be made transparent to the customer.
 If data is damaged, everything is damaged and
lost.
 Hence, care should be taken to protect the
infrastructure.
Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 39
Cloud Computing: Security Aspects
3. Storage Security:
 It is important to be informed of who owns the
data and the location where it is stored.
 Data leak, snooping, malware attacks, etc. are
all threats to the stored data and can be listed
under storage security.
 Appropriate antivirus software and periodic
monitoring, should help protect the data.
4. Network Security:
 Data is stored in the cloud via the Internet,
and hence all network threats become a
possibility.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 40

Summary
 Cloud computing has become one of the most used technology components in modern day
applications, which not only provides storage but also supports data analytics.
 Cloud services could be any one of the following:
 Software as a Service (SaaS): Complete software application as a service is provided to the user.
 Platform as a Service (PaaS): Development tools, APIs, libraries, etc. will be divided by the cloud service
provider. User have to build, manage and maintain the applications.
 Infrastructure as a Service (IaaS): User should be provided with virtual machine support, where the user
does not need to know and worry about the infrastructure. Everything should be taken care by the service
provider. User will manage the machines, select the OS and underlying applications.
 The three deployment models generally used for public, private and hybrid
 Private cloud deployment model can be opted wherever confidentiality matters the most.
 When its come to public cloud deployment model, the cloud service provider owns all the
resource which include hardware/infrastructure and software. Cloud service provider will take
care of all resource management.

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 41

Summary
 Hybrid development is a mix of both public and private deployment model. In this approach the
resource offered and manage are both in-house and third party based.
 There are many challenges one could face while opting for cloud storage with IoT applications
some of these are as follows:
 Privacy and security
 Bandwidth cost
 Migration and portability
 Reliability and availability
 Costing
 Data ownership
 Expertise
 Selecting a CSP is not easy. Many parameters are to be considered before choosing the best
option.
 With fog computing, it become possible to analyze the data at a place closer to where it is
generated. 
Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 42
Summary
 Fog computing provides the following advantages:
 Minimal amount of data send to cloud
 Reduce bandwidth consumption
 Reduce data latency
 Improve data security
 Immediate processing of data

Prof. Kalpesh H Surati #3160716 (IoT)  Unit 4 – Cloud for IoT 43