CSE 2032

Introduction to
Fog Computing

Module IV
 Content
❖ Fog Integration Technologies ❖ Fog-IoT:
with IoT ✔ Architectural model,

✔ Security Management for Edge Cloud Architectures ✔ Challenges on IoT Stack Model via
TCP/IP Architecture,
✔ Fog Computing Realization for Big Data Analytics
✔ Data Management,
✔ Data Analytics in the Fog, Prototypes and Evaluation
✔ Security and privacy issues
❖ Case Study:
❖ Intelligent Traffic Lights Management (ITLM) System

❖ Fog Computing for Smart Transportation Applications

 Security Management for Edge Cloud
Architectures
❖ Edge Cloud Architectures
represent a distributed
computing paradigm.
❖ Unlike traditional cloud
computing centralized in data
centers, edge architectures bring
computing resources closer to
the data source. (Decentralized)
 Security Management for Edge Cloud
Architectures
❖ Decentralization: Edge clouds
distribute computational
capabilities across a network's
edge.

❖ Proximity: Data processing

occurs closer to where it's
generated, reducing latency.

❖ Real-time Processing:
Suitable for applications requiring
instant responses, such as IoT and
AI.
 Security Management for Edge Cloud
Architectures
Components of Edge Clouds

❖ Edge Nodes: Devices at the

periphery of the network (e.g.,
routers, IoT devices, edge servers).

❖ Fog Nodes: Intermediate nodes

between edge devices and the
central cloud.
 Security Management for Edge Cloud
Architectures
❖ Edge Cloud security is a combination of tools and practices that protect edge
infrastructure and workloads in remote locations.
❖ It protects data that is generated and processed from an organization's on-premises data
center or public cloud. It includes

✔ Visibility: Administrators should be able to see the entire network.

✔ Encryption: Data should be encrypted at rest and in transit.
✔ Monitoring: Automated monitoring tools should be in place.
✔ Access: Access to alter data and network resources should be restricted.
 Security Management for Edge Cloud
Architectures
❖ Edge Cloud security can also:
✔ Address DDoS attacks in real-time.
✔ Protect data that enters and exits the network.
❖ A Distributed Denial of Service (DDoS) attack is a malicious attempt to disrupt the normal
functioning of a network, service, or website by overwhelming it with a flood of internet
traffic.
❖ In a DDoS attack, multiple compromised computers or devices are used to generate and
send an unusually high volume of traffic to a target, causing it to become slow,
unresponsive, or completely unavailable.
 Security Management for Edge Cloud
Architectures
❖ Addressing Distributed Denial of Service (DDoS) attacks in real-time is crucial to
minimize their impact and ensure the availability of services.

1. Traffic Filtering and Rate Limiting:

✔ Implement traffic filtering and rate limiting mechanisms at the network perimeter to
identify and block malicious traffic.
✔ Use intrusion prevention systems (IPS) to detect and mitigate DDoS attacks by blocking
or limiting the rate of incoming traffic.
 Security Management for Edge Cloud
Architectures
2. Content Delivery Networks (CDNs):
✔ Distribute content across a network of geographically dispersed servers using a CDN. This
helps absorb and distribute the impact of DDoS attacks.
✔ CDNs can cache static content and provide load balancing to reduce the load on the origin
server during an attack.

3. Anycast DNS:
✔ Use Anycast DNS to distribute DNS resolution requests across multiple servers in different
locations.
✔ This helps distribute the load and mitigate DDoS attacks targeting DNS infrastructure.
 Security Management for Edge Cloud
Architectures
4. Web Application Firewalls
(WAF):
✔ Employ WAFs to inspect and filter HTTP
traffic, blocking malicious requests and
payloads that are indicative of DDoS
attacks.
✔ Configure WAFs to detect and mitigate
common DDoS attack vectors, such as
HTTP floods and application layer
attacks.
 Questions
1. How would you design security policies for edge cloud architectures to
ensure data integrity and confidentiality?
2. How do you address the unique security challenges posed by the
distributed nature of edge cloud architectures?
3. How would you implement secure device onboarding and
authentication in edge cloud architectures?
4. How do you manage security updates and patches in a large-scale
edge cloud deployment?
 Answers
1. How would you design security policies for edge cloud architectures to
ensure data integrity and confidentiality?
❖ Security policies for edge cloud architectures should include
✔ Encryption protocols for data in transit
✔ Access control mechanisms, and regular security audits.
✔ Implementing secure communication channels, such as TLS/SSL, and
using encryption algorithms for sensitive data helps maintain data integrity
and confidentiality.
 Answers
2. How do you address the unique security challenges posed by the
distributed nature of edge cloud architectures?
❖The distributed nature of edge cloud architectures requires a
comprehensive approach. This includes
✔ Deploying security measures at each edge node, implementing
secure communication protocols, and leveraging edge-specific
security solutions.
✔ Additionally, centralized security management tools can be used to
monitor and respond to security events across the distributed edge
infrastructure.
 Answers
3. How would you implement secure device onboarding and authentication
in edge cloud architectures?

✔ Secure device onboarding involves verifying the identity of

devices before granting access.
✔ This can be achieved through mutual authentication, where both
the device and the edge node authenticate each other.
✔ Using secure boot processes, device certificates, and
multi-factor authentication are common practices in ensuring
secure device onboarding.
 Answers
4. How do you manage security updates and patches in a large-scale
edge cloud deployment?

✔ Managing security updates requires a systematic approach.

✔ This includes using automated patch management systems,
scheduling updates during low-traffic periods, and testing
updates in a controlled environment before deployment.
✔ Additionally, having a rollback plan in case of issues ensures
minimal disruption.
Fog Computing Realization for Big Data
Analytics
❖ Big Data Analytics refers to the process of
examining and interpreting large and
complex datasets to extract meaningful
insights, patterns, and trends.
❖ It involves the use of advanced analytics
techniques to analyze data that is too
voluminous, varied, or fast-moving for
traditional data processing systems.
 Fog Computing Realization for Big Data
Analytics
Characteristics of Big Data:
• Volume: Deals with the sheer amount of data generated,
often in terabytes or petabytes.
• Velocity: Refers to the speed at which data is generated,
processed, and analyzed in real-time.
• Variety: Encompasses diverse types of data, including
structured, unstructured, and semi-structured data from
various sources.
• Veracity: Addresses the quality and reliability of data,
considering the accuracy and trustworthiness of
information.
 Fog Computing Realization for Big Data
Analytics
Key Components:
•Data Collection: Involves gathering data from various sources, such as sensors,
social media, and transaction records.
•Data Storage: Utilizes distributed storage systems capable of handling large-scale
data, often stored in a Hadoop Distributed File System (HDFS) or cloud-based
storage.
•Data Processing: Involves the use of distributed computing frameworks like
Apache Spark or Hadoop MapReduce to process and analyze data in parallel
across multiple nodes.
•Data Analysis: Utilizes statistical algorithms, machine learning, and data mining
 Fog Computing Realization for Big Data
Analytics
Analytical Tools:
•Hadoop: A framework for distributed storage and processing of large datasets.

•Spark: An open-source, distributed computing system for big data processing and
analytics.
•NoSQL Databases: Such as MongoDB, Cassandra, and Couchbase, which are used
for storing and retrieving unstructured or semi-structured data.
•Machine Learning Libraries: Like TensorFlow and scikit-learn, used for building
predictive models.
•Data Visualization Tools: Such as Tableau, Power BI, and D3.js, for presenting
insights in a visually understandable format.
 Questions
1. How does the shift from centralized cloud analytics to edge analytics in
fog computing impact the design and implementation of big data
analytics algorithms?
2. How can big data analytics algorithms be optimized to efficiently utilize
the limited computational resources available in fog computing
environments, considering the constraints of edge devices?
3. How do considerations for data movement and storage change when
implementing big data analytics at the edge in fog computing, and
what impact does this have on latency and bandwidth?
 Answers
1. How does the shift from centralized cloud analytics to edge analytics in
fog computing impact the design and implementation of big data analytics
algorithms?
❖ Edge analytics introduces challenges related to
✔ Resource constraints,
✔ Data variability, and
✔ The need for lightweight algorithms that can operate efficiently on edge
devices.
 Answers
2. How can big data analytics algorithms be optimized to efficiently utilize
the limited computational resources available in fog computing
environments, considering the constraints of edge devices?

❖ Big data analytics algorithms in fog computing must be optimized for

resource-efficient execution, considering the limited computational power
and energy constraints of edge devices.
 Answers
3. How do considerations for data movement and storage change when
implementing big data analytics at the edge in fog computing, and what
impact does this have on latency and bandwidth?

❖ Minimizing unnecessary data movement and leveraging edge storage can

help mitigate latency and bandwidth challenges in fog computing
environments.
Fog Computing Realization for Big Data
Analytics
• Internets of Things (IoT) deployments generate large quantities of data
that need to be processed and analyzed in real time.
• Current IoT systems do not enable low-latency and high-speed processing
of data and require offloading data processing to the cloud.
• Although cloud computing optimizes resource utilization, it does not
provide an effective solution for hosting big data applications.
Fog Computing Realization for Big Data
Analytics
Issues in adopting IoT-driven services,
1. Moving large amounts of data over the nodes of a virtualized computing
platform may incur significant overhead in terms of time, throughput,
energy consumption, and cost.
2. The cloud may be physically located in a distant data center, so it may
not be possible to service IoT with reasonable latency and throughput.
3. Processing large quantities of IoT data in real time will increase as a
proportion of workloads in data centers, leaving providers facing new
security, capacity, and analytics challenges.
Fog Computing Realization for Big Data
Analytics
Issues in adopting IoT-driven services
4. Current cloud solutions lack the capability to accommodate analytic
engines for efficiently processing big data.
5. Existing IoT development platforms are vertically fragmented. Thus, IoT
innovators must navigate between heterogeneous hardware and
software services that do not always integrate well together
Fog Computing Realization for Big Data
Analytics

Typical data analytics flow

Fog Computing Realization for Big Data
Analytics
Solution
Fog Computing Realization for Big Data
Analytics
Solution
❖ Fog-engine is a customizable and agile heterogeneous platform that is
integrated to an IoT device.
❖ The Fog-engine allows data processing in the cloud and in the
distributed grid of connected IoT devices located at the network edge.
❖ It provides facilities for offloading data and interacting with the cloud as a
gateway.
❖ A gateway enables devices that are not directly connected to the
Internet to reach cloud services.
Fog Computing Realization for Big Data
Analytics
Projects

Real-time Sentiment Analysis:

▪ Analyze social media feeds or customer reviews in real-time to determine sentiment

towards a product, service, or brand.

▪ Use Natural Language Processing (NLP) techniques to classify and analyze text data.

Social Network Analysis:

▪ Analyze social network data in real-time to identify trends, influential users, and potential
viral content.

▪ Use graph analytics and machine learning to extract insights from social media data.
Fog Computing Realization for Big Data
Analytics
Real-time sentiment analysis involves analyzing text data in real-time to
determine the sentiment or emotional tone expressed.
Objectives: To understand customer opinions, feedback, and reactions on
social media, reviews, or other text sources. - To improve Business
1.Data Collection:
▪ Set up a data pipeline to collect real-time data from social media
platforms, customer reviews, or any other relevant sources.
▪ Use APIs or web scraping techniques to continuously fetch new data.
Fog Computing Realization for Big Data
Analytics
2. Data Preprocessing:
▪ Clean and preprocess the text data by removing noise, handling
missing values, and standardizing the format.
▪ Tokenize the text into words or phrases for analysis.
3. Sentiment Analysis Model:
▪ Choose or build a sentiment analysis model. You can use pre-trained
models or train your own model using sample data.
▪ Fine-tune the model on a sentiment analysis dataset to adapt it to your
specific domain.
Fog Computing Realization for Big Data
Analytics
4. Real-time Analysis:
▪ Implement a real-time analysis component that continuously processes
incoming data.
▪ Use the trained model to predict sentiment for each piece of incoming
text in real-time.
5. Visualization:
▪ Create a dashboard or visualization tool to display real-time sentiment
trends.
▪ Include features such as sentiment distribution, sentiment over time,
and other relevant visualizations.
Fog Computing Realization for Big Data
Analytics
6. Alerts and Notifications:
▪ Implement an alerting system to notify stakeholders when there is a
significant shift in sentiment.
▪ Set up thresholds for positive or negative sentiment and trigger alerts
accordingly.
7. Feedback Loop:
▪ Implement a feedback loop to continuously improve the model.
▪ Allow users to provide feedback on the accuracy of sentiment
predictions, and use this feedback to retrain the model periodically.
 Internet of Things
❖The internet of things, or IoT, is a system of interrelated
computing devices, mechanical and digital machines,
objects, animals or people that are provided with
unique identifiers (UIDs) and the ability to transfer data
over a network without requiring human-to-human or
human-to-computer interaction.

❖The Internet of Things (IoT) describes the network of

physical objects—“things”—that are embedded with
sensors, software, and other technologies for the
purpose of connecting and exchanging data with other
devices and systems over the internet.
36
How its Works

37
 Sensors:
❖ A sensor or device is an important
component for gathering live data from
the surrounding environment.
❖ The nature of this data can vary.
❖ This could be as simple as your phone
having a temperature sensor, GPS, an
accelerometer, or as complex as a live
video feature on a social media platform.
❖ Sensors make it possible for IoT devices
to connect to the real world and
environment.

38
 Connectivity
❖ Upon collection, all data is sent to a cloud
infrastructure.
❖ This could be done by connecting the sensors
to the cloud using a variety of communication
mediums such as mobile or satellite networks,
Bluetooth, WI-FI, WAN, etc.
❖ Various IoT devices use different types of
connectivity.

39
 Data Processing:

Data Processing:
❖ Once the data has been
collected, and has reached the
cloud, it is the responsibility of the
data processors to process it.
❖ Data processing software can
enhance IoT devices in a wide
range of ways, from adjusting the
temperature of the air conditioner
to recognizing faces on mobile
phones.
40
 User Interface

• The information needs to be available to the

end-user in some way which can be achieved
by triggering alarms on their phones or
sending them notification through email or text
message.
o The user sometimes might need an interface
which actively checks their IoT system.
• For example, the user has a camera installed
in his home. He wants to access video
recording and all the feeds with the help of a
web server.
41
 IoT benefits to organizations
✔ monitor their overall business processes;
✔ improve the customer experience (CX);
✔ save time and money;
✔ enhance employee productivity;
✔ integrate and adapt business models;
✔ make better business decisions; and
✔ generate more revenue.

42
 Why we need Internet of Things (IoT)
• Over the past few years, IoT has become one of the most important
technologies of the 21st century.
• Now that we can connect everyday objects—kitchen appliances, cars,
thermostats, baby monitors—to the internet via embedded devices’
• So seamless communication is possible between people, processes, and
things.
• By means of low-cost computing, the cloud, big data, analytics, and mobile
technologies, physical things can share and collect data with minimal human
intervention.

43
 Advantages of IoT
• Ability to access information from anywhere at any time on any device;
• Improved communication between connected electronic devices;
• Transferring data packets over a connected network saving time and
money; and
• Automating tasks helping to improve the quality of a business's
services and reducing the need for human intervention.

44
 Disadvantages IoT
• As the number of connected devices increases and more information
is shared between devices, the potential that a hacker could steal
confidential information also increases.
• Enterprises may eventually have to deal with massive numbers --
maybe even millions -- of IoT devices, and collecting and managing
the data from all those devices will be challenging.
• If there's a bug in the system, it's likely that every connected device
will become corrupted.
• Since there's no international standard of compatibility for IoT, it's
difficult for devices from different manufacturers to communicate with
each other.
45
Key Features

46
 Key Features
AI − IoT essentially makes virtually anything “smart”, meaning it enhances
every aspect of life with the power of data collection, artificial intelligence
algorithms, and networks. This can mean something as simple as enhancing
your refrigerator and cabinets to detect when milk and your favourite cereal
run low, and to then place an order with your preferred grocer.
Connectivity − New enabling technologies for networking, and specifically IoT
networking, mean networks are no longer exclusively tied to major providers.
Networks can exist on a much smaller and cheaper scale while still being
practical. IoT creates these small networks between its system devices.
Sensors − IoT loses its distinction without sensors. They act as defining
instruments which transform IoT from a standard passive network of devices
into an active system capable of real-world integration. 47
Active Engagement − Much of today's interaction with connected technology
happens through passive engagement. IoT introduces a new paradigm for
active content, product, or service engagement.
Small Devices − Devices, as predicted, have become smaller, cheaper, and
more powerful over time. IoT exploits purpose-built small devices to deliver
services.

2022: Bosch Links IoT with AI (Bosch Center for Artificial

Intelligence.)Bosch is focusing on the connection of IoT with AI,
basically tying the Internet of Things and artificial intelligence
together. (Lot of Scope for Final Year Experts in AI
48
 Questions
Which of the following is not a key component of an IoT system?
a. Sensors
b. Actuators
c. Blockchain
d. Connectivity
 Questions
What is the primary purpose of IoT devices?
a. Entertainment

b. Monitoring and Control

c. Social Networking
d. Gaming
 Questions
Which communication protocol is commonly used for IoT devices due to its lightweight
nature?
a. HTTP
b. MQTT
c. TCP
d. UDP
 Questions
Which of the following is a challenge in IoT security?
a. Lack of standardization
b. Overabundance of connectivity options
c. Limited data storage capacity
d. Slow data transmission speeds
 Questions
What is the purpose of Edge Computing in IoT?
a. To centralize all IoT devices in one location
b. To process data closer to the source rather than relying on a centralized cloud
c. To increase the number of IoT devices in a network
d. To improve the aesthetics of IoT devices
 Questions
1. How can the industry address the challenge of interoperability among
different IoT devices and platforms?
2. What role should standardization play in ensuring a seamless and
integrated IoT ecosystem?
3. How can IoT devices be designed to prioritize the user experience and
address the needs of diverse user groups?
4. What considerations should be made to ensure that IoT technology
enhances human well-being rather than detracts from it?
 Answers
1. How can the industry address the challenge of interoperability among
different IoT devices and platforms?

❖ Embracing common communication protocols and supporting interoperability

standards.
 Answers
2. What role should standardization play in ensuring a seamless and
integrated IoT ecosystem?
❖ Facilitating collaboration, reducing fragmentation, and improving overall
system reliability.
 Answers
3. How can IoT devices be designed to prioritize the user experience and
address the needs of diverse user groups?
❖ Involving users in the design process, considering accessibility, and focusing
on simplicity and intuitiveness.
 Answers
4. What considerations should be made to ensure that IoT technology
enhances human well-being rather than detracts from it?

❖ Ethical design principles, user feedback loops, and continuous evaluation of

societal impacts.
 Test
1. What technology is often used to implement network functions in network slicing?
2.
(1 Marks)

2. What is Massive Machine Type Communications (mMTC). (1 Marks)

3. What are the primary components and services provided by fog and edge
computing middleware, and how do they facilitate distributed computing? (2 Marks)
4. Write short notes on Network Slicing in Software-Defined Clouds.(3 Marks)
5. Explain about Middleware for Fog and Edge Computing. .(3 Marks)
Fog-IoT Architectural Model
❖ A reference framework of fog-ioT computing is designed comprising
of the seven levels namely
1. level 1: virtualized and physical
2. level 2: fog devices, servers, and gateway
3. level 3: monitoring
4. level 4: preprocessing and post-processing
5. level 5: storage and resource management
6. level 6: security and
7. level 7: application
Fog-IoT Architectural Model
 Fog-IoT Architectural Model
❖ The goal of these levels is to work together to push a task for execution
from an IoT to fog nodes and then to the cloud.
❖ These levels focus on carrying different tasks such as
✔ information management,
✔ data analysis, processing of data,
✔ categorizing of information to cloud servers and fog servers, and
✔ various other tasks based on services of the fog and cloud, and the
demand of applications from the users.
 Fog-IoT Architectural Model
Level 1: Physical and virtual sensors:
❖ The different forms of information produced by the sensors are the
basic information generator of Fog Computing.
❖ This information could be produced from different devices such as
intelligent homes and devices, surveillance systems of CCTV and
traffic, automated driving vehicles, humidity and temperature sensors,
and so on.
❖ The physical level contains virtual and physical sensors, where any
information producing device may fall into any of these clusters.
 Fog-IoT Architectural Model
Level 2: Fog device, server and gateway
❖ An IoT or an independent device may be a fog server, fog device, or a
gateway
❖ it is noticeable that the fog server must have a better configuration than
the fog gateway and devices since it controls numerous fog devices;
❖ To make fog server run many factors are involved like configuration of
hardware, network connectivity, devices it can control, etc
❖ The fog server must have better computation and storage capability with
fog devices. It should also have advanced computation and storage
Fog-IoT Architectural Model
Level 3: Monitoring
❖ The performance of the system and resources are kept track by the
monitoring level along with utility and feedback
❖ Based on the network load and availability of the resources the
prediction monitor can signal the performance of the fog system.
❖ This is required as it is used to maintain the required QoS attributes in
SLA’s.
 Fog-IoT Architectural Model
Level 4: Pre and post processing
❖ This level works on data analysis of basic and advanced data, it has
multiple components
❖ In this level, it plays the role of obtaining the data by analyzing, filtering,
trimming, and reconstructing the data as and when required
❖ Once the data is processed, the component named data flow selects
where the data should be stored whether locally at the fog or in the cloud
for storage for a long duration.
 Fog-IoT Architectural Model
Level 5: Storage and resource management
❖ The storage module is in charge of storing data using storage
virtualization.
❖ The component named data backup is responsible to ensure the data
availability and data loss.
❖ The concept of storage virtualization contains a pool of devices that are
responsible for storage in a network that acts as an individual storage
device.
❖ The main benefit of storage virtualization is that the cost of hardware and
 Fog-IoT Architectural Model
Level 6: Security
❖ The security level maintains all the issues related to security like communication
encryption and protected information storage.
❖ To avoid intrusion of malicious users it is important to maintain security by having
encryption between different communications.
❖ Encryption component will encrypt various connections to and fro from the IoT
devices and the cloud.
❖ Most of the fog components are connected through a wireless connection, so it is
important to maintain security.
 Fog-IoT Architectural Model
Level 7: Application
❖ Almost all applications that have latency as an issue started to take advantage of
the fog environment
❖ These included any type of utility services that could integrate with fog to provide
better service and minimize cost. In an application, in which the system uses
Augmented Reality can adopt fog infrastructure as it will change the current world
in the future.
❖ The requirements of processing in real-time using augmented reality can be
catered by fog environment which can cause prolonged improvement in many
services of augmented reality.