CSE 2032

Introduction to
Fog Computing
 Module I
AFTER GOING THROUGH THIS MODULE YOU SHOULD BE ABLE TO:

1. DESCRIBE ABOUT FOG COMPUTING TECHNOLOGY

2. IDENTIFY THE DIFFERENCES BETWEEN FOG AND CLOUD COMPUTING

3. KNOW ABOUT ARCHITECTURE, ADVANTAGES, APPLICATIONS AND CHALLENGES ASSOCIATED WITH

FOG

4. DESCRIBE ABOUT EDGE COMPUTING

5. DIFFERENTIATE BETWEEN CLOUD, FOG AND EDGE COMPUTING

6. COMPARE DIFFERENT APPLICATIONS OF EDGE COMPUTING

PRESIDENCY UNIVERSITY, BANGALORE 2

 Module I –L2
LO1: DEFINE COMPUTER NETWORKS
LO2:CLASSIFY VARIOUS INTERCONNECTING
DEVICES
Introduction
Basics
A computer network is a collection of
interconnected devices, such as computers,
servers, routers, switches, and other
hardware components, that are linked
together to facilitate communication, data
sharing, and resource sharing.
These devices are connected through
various Interconnecting Devices.
Introduction
Interconnecting Devices
1. Hub: Simple device that broadcasts data to all
connected devices, operating at Layer 1
(physical layer).
2. Switch: Intelligent device that forwards data
only to the intended recipient based on MAC
addresses, operating at Layer 2 (data link layer).
3. Router: Device that routes data between
different networks and provides network
segmentation and security, operating at Layer 3
(network layer).
 Introduction
IP Addressing
❖ An IP address represents a unique address that
distinguishes any device on the internet or any network
from another.
❖ An IP address is the identifier that enables your device to IPv4
send or receive data packets across the internet.
❖ It holds information related to your location and therefore
making devices available for two-way communication.
1. IPv4 64 Bit Address)
2. IPv5 IPv6
3. IPv6 128 Bit Address)

PRESIDENCY UNIVERSITY, BANGALORE 6

 Traditional Networks
❑ Physical Infrastructure: Infrastructure is located
on-site, often requiring dedicated server rooms
or data centers.

❑ Cost Structure: High upfront costs are associated

with purchasing and maintaining hardware and
infrastructure
❑ Scalability: Scaling up requires purchasing new
hardware, which might involve delays and
additional costs.
❑ Scaling down can lead to underutilized
resources.

PRESIDENCY UNIVERSITY, BANGALORE

 Traditional Networks
❑ Maintenance: IT teams are responsible for managing hardware,
software updates, security patches, and troubleshooting.

❑ Accessibility: Access to resources is often limited to the physical

location of the network.

❑ Disaster Recovery: Organizations need to implement their own

disaster recovery plans, including data backup and redundancy
solutions.

PRESIDENCY UNIVERSITY, BANGALORE

Module I –L3
LO1: Identify IP Address notation

LO2:Differentiate Traditional Networks and Cloud Computing

 Cloud Computing
Delivery of computing services such as servers, storage, databases,
networking, software, analytics, intelligence, and more, over the Internet

Cloud Computing provides an alternative to the on-premises

datacenters

PRESIDENCY UNIVERSITY, BANGALORE 10

 Cloud Computing

PRESIDENCY UNIVERSITY, BANGALORE 11

 Cloud Computing

PRESIDENCY UNIVERSITY, BANGALORE 12

 Cloud Computing
❑ AMAZON
Compute Services:
o Amazon EC2 : virtual servers for running applications
and workloads.
o AWS Lambda: to run code without provisioning or
managing servers, suitable for event-driven
applications.
❑ RACKSPACE
o To help businesses with their cloud infrastructure,
management, and support needs.

PRESIDENCY UNIVERSITY, BANGALORE 13

 Cloud Computing
❑ Microsoft Azure
o Compute Services: Azure Virtual Machines (virtualized
Windows or Linux servers) Azure App Service (building
and hosting web applications)
o Storage Services: Azure Blob Storage
o Database Services: Azure SQL Database, ySQL and
PostgreSQL
o Networking Services: Azure Virtual Network:
o AI and Machine Learning: Azure Machine Learning
o Internet of Things (IoT): Azure IoT Hub - Managed
service for connecting, monitoring, and managing IoT
PRESIDENCY UNIVERSITY, BANGALORE 14
 Cloud Computing
❑ IBM CLOUD
o IBM Virtual Servers
o IBM Cloud File Storage- Network-attached storage
o MySQL, and MongoDB
o IBM Virtual Private Cloud - Network environments
o IBM Watson Studio – AI Models
o IBM Continuous Delivery: DevOps toolchain for automating
application building, testing, and deployment.

PRESIDENCY UNIVERSITY, BANGALORE 15

 Cloud Computing
❑ Google Cloud Platform (GCP)
o Compute Services: Google App Engine: Platform as a Service
(PaaS) for building and deploying applications without managing
the underlying infrastructure.
o Storage Services: Unlimited Storage
o Database Services: NoSQL database for building web, mobile,
and server applications.
o Networking Services: Google Virtual Private Cloud
o AI and Machine Learning: Google Cloud AI Platform
o Internet of Things (IoT): Google Cloud IoT Core

PRESIDENCY UNIVERSITY, BANGALORE 16

Traditional Networks Vs Cloud Computing

Aspect Traditional Networking Cloud Computing

On-premises hardware and Virtualized infrastructure in
Infrastructure
networking data centers
Scaling may require manual Easily scalable through
Scalability
hardware cloud providers
Manual provisioning and Automated provisioning and
Resource Management
management management
Operating expenditure
Cost Model Capital expenditure (CapEx)
(OpEx)
Traditional Networks Vs Cloud Computing
Aspect Traditional Networking Cloud Computing
Higher upfront hardware Lower upfront costs,
Initial Investment
and setup costs pay-as-you-go model
Requires in-house IT Managed by cloud
Maintenance
teams provider
Location Dependency Tied to physical location Accessible from anywhere
Typically manually Built-in redundancy for
Redundancy
implemented high availability
Limited by hardware and Highly flexible and
Flexibility
capacity adaptable
Traditional Networks Vs Cloud Computing
Aspect Traditional Networking Cloud Computing
Relatively slower Rapid deployment of
Speed of Deployment
provisioning resources
Manual backup and Built-in disaster recovery
Disaster Recovery
recovery processes mechanisms
Mostly managed by the Shared responsibility with
Security Responsibility
organization the provider
Manual software and Automatic updates and
Updates and Upgrades
hardware updates upgrades
Hardware maintenance, Centralized data centers'
Environmental Impact
power, cooling impact analyzed
 Why Fog Computing?

Example:
”Imagine a city with a complex
network of traffic signals, surveillance
cameras, and sensors installed at
various intersections.
Goal:
To optimize traffic flow, reduce
congestion, and improve overall road
safety”.

PRESIDENCY UNIVERSITY, BANGALORE 22

 Why Fog Computing?

Cloud Computing Approach

❖ In a traditional cloud-centric approach, all the data
collected from cameras, sensors, and traffic signals
would be sent to a centralized cloud server for
processing and analysis.
❖ This server would then generate recommendations for
adjusting traffic light timings and other control
measures based on the collected data.
❖ However, this approach may introduce latency due to
the time it takes to transmit data to the cloud and
receive processing results back, potentially leading to
suboptimal traffic management decisions.

PRESIDENCY UNIVERSITY, BANGALORE 23

 Why Fog Computing?

Fog Computing Approach

▪ At each traffic intersection, fog nodes are

installed.
▪ These fog nodes have processing power,
memory, and connectivity capabilities.
▪ They can analyze data locally in real-time
and make immediate decisions.

PRESIDENCY UNIVERSITY, BANGALORE 24

 Decentralized computing architecture that extends the
Fog Computing capabilities of cloud computing to the edge of the
network.

It involves the distribution of computing

resources and services closer to the data source,
user devices, and edge devices in a hierarchical
manner.
Fog computing enables data processing at the
edge of the network, reducing the need to send
all data to a centralized cloud. This leads to faster
response times and reduced latency.

PRESIDENCY UNIVERSITY, BANGALORE 25

Fog Computing

The goal of fog computing is to reduce

latency, enhance real-time data processing,
and improve overall efficiency in various
applications, particularly those in the realm
of the Internet of Things (IoT) and
real-time data analytics.

PRESIDENCY UNIVERSITY, BANGALORE 26

Fog Computing
In fog computing, data is processed, analyzed, and
acted upon at or near the point of origin, rather than
being sent to a centralized cloud server for processing.

This approach is particularly beneficial for applications

that require rapid responses, low latency, and reduced
reliance on constant internet connectivity

Note
"Fog computing empowers devices to think globally
and act locally, ushering in a new era of decentralized
intelligence."
PRESIDENCY UNIVERSITY, BANGALORE 27
Module I –L4
LO1: Outline the Characteristics of Fog Computing

LO2:Determine the critical issues of Fog Nodes

 Fog Computing Characteristics
Proximity to Data Source:

✔ Fog computing places computing

resources closer to where data is
generated, typically at the edge of the
network
Example

Sensors, IoT devices, and

data-producing machines
 Fog Computing Characteristics
Low Latency

✔ By processing data locally, fog

computing reduces the time it takes
for data to travel to a distant cloud
server and back.
This enables quicker response times, which
is crucial for time-sensitive applications.
 Fog Computing Characteristics
Real-Time Decision-Making

✔ Fog computing allows immediate analysis and decision-making at the edge,

making it suitable for applications where real-time processing and quick
responses are essential, such as autonomous vehicles, industrial automation,
and smart cities.

Bandwidth Efficiency:

✔ Fog computing reduces the amount of data that needs to be transmitted to the
cloud for processing, which optimizes network bandwidth and reduces
congestion.
 Fog Computing Characteristics
Resilience: Distributed Architecture:

✔Fog computing systems can continue to operate ✔Fog computing involves a network of
even if there are disruptions in internet connectivity.
edge devices and gateways that work
Local processing ensures that critical functions can
still be performed without relying on a constant together to process and manage
connection to the cloud. data.

Privacy and Security: ✔This distributed architecture allows

for scalability and fault tolerance.
✔Data can be processed locally, minimizing the need
to transmit sensitive information over the internet,
which can enhance data privacy and security.
 Critical Issues of Fog Nodes
Resource Constraints:

✔Fog nodes are often resource-constrained devices, such as IoT sensors,

gateways, or edge servers. These devices may have limited processing power,
memory, and storage capacity.

✔Heterogeneity: :

✔Fog computing environments typically consist of a wide range of heterogeneous

devices with varying capabilities, operating systems, and communication
protocols. Ensuring interoperability and efficient communication among these
devices can be a challenge.
 Critical Issues of Fog Nodes
Scalability:

✔As the number of fog nodes in a network grows, managing and scaling the
infrastructure becomes complex. Ensuring that the system can handle additional
nodes while maintaining performance and reliability is a significant challenge.

Energy Efficiency:

✔Many fog nodes are battery-powered or have limited access to power sources.
Optimizing energy consumption while performing computation and
communication tasks is critical for the longevity of these devices.
 Critical Issues of Fog Nodes
Quality of Service (QoS):

✔Fog computing aims to provide low-latency, high-throughput services close to

the data source. Ensuring consistent QoS in a dynamic and distributed fog
environment is a challenge, especially when dealing with variable network
conditions.

✔Security:

✔Fog nodes are often deployed in physically vulnerable or untrusted

environments. Ensuring the security of data, applications, and communication
between nodes is crucial. Fog nodes can be vulnerable to physical attacks,
software vulnerabilities, and network attacks.
 Critical Issues of Fog Nodes
Data Management:

Managing and processing data at the edge can be complex. Deciding which data
to process locally on fog nodes and which data to send to the cloud for further
analysis requires careful planning. Data storage, retention, and synchronization
challenges also arise.

Cost Management:

✔Deploying and maintaining a large number of fog nodes can be costly. Balancing
the benefits of fog computing with the associated expenses is a constant
challenge for organizations.
Module I –L5
LO1: Identify the need for Fog Computing in IoT

LO2:Determine the challenges of IoT devices

Fog Computing in IoT
Fog Computing in IoT
Module I –L6
LO1: Identify the deployment strategies of Fog

LO2:Differentiate between public, private and hybrid models

 Fog Deployment Models
Centralized Fog Model:
∙ Fog nodes are placed closer to the edge of the
network but are still managed by a central cloud
infrastructure.
∙ Fog nodes collect data, perform preliminary
processing, and then send relevant information to
the central cloud for further analysis and storage.

∙ This model is useful when there's a need for

centralized control and data analysis, while still
benefiting from reduced latency at the edge.
 Fog Deployment Models
Distributed Fog Model:
▪ Fog nodes are distributed across various locations
and are responsible for local processing and
decision-making.

▪ Each fog node operates autonomously, processing

data from nearby edge devices

▪ This model is suitable for applications that require

localized decision-making, efficient resource
utilization, and resilience.
 Fog Deployment Models
Hierarchical Fog Model:
∙ Combines centralized and distributed approaches.

∙ Fog nodes are organized into hierarchical tiers,

with each tier responsible for specific processing
tasks.

∙ Data is processed at different levels of the

hierarchy: local nodes process immediate data,
intermediary nodes perform more complex
analysis, and a central cloud might perform global
analytics.

∙ This model balances local processing with

centralized coordination and scalability.
 Fog Deployment Models
Mesh Fog Model
▪ Fog nodes are interconnected in a mesh network topology.

▪ Each fog node can communicate with neighboring nodes,

facilitating collaborative processing and fault tolerance.

▪ Data can be dynamically routed through the mesh, allowing for

load balancing and efficient data flows.

▪ This model suits scenarios where data needs to be shared and

processed across multiple fog nodes for collaborative tasks.
 Fog Deployment Models
Mobile Edge Computing (MEC) Model:
▪ MEC is a specific fog deployment model focused on mobile
networks.

▪ Fog nodes are deployed at the edge of cellular networks, providing

services to mobile users with reduced latency.

▪ This model enhances mobile applications with real-time

capabilities and enables network operators to offer services at the
edge.
Module I –L7
LO1: List down the application domain of Fog Computing

LO2:Explain Remote Monitoring Using Fog Computing

 APPLICATIONS OF FOG
Smart Cities
 APPLICATIONS OF FOG
Smart Car and Traffic Control System
 APPLICATIONS OF FOG
Smart Grids
 APPLICATIONS OF FOG
Smart Healthcare Systems
Module I –L8
LO1: Define Inner Edge Device.

LO2:Differentiate Outer Edge and inner Edge Device

 Hierarchy of Fog and Edge Computing
▪ Fog and edge computing are both paradigms that extend cloud computing
capabilities

▪ Placed closer to the data source or end-users to reduce latency, improve

responsiveness, and enable real-time processing

▪ However, they operate at different proximity levels to the end-users and

have distinct characteristics
 Hierarchy of Fog and Edge Computing
▪ FEC servers deployed at three
edge layers—

1) Inner-edge,

2) Middle-edge, and

3) Outer-edge
 Hierarchy of Fog and Edge Computing
Inner-edge:

▪ Inner-edge also known as near-the-edge corresponds to countrywide, statewide

and regional Wide Area Network (WAN) of enterprises, Internet Service
Providers (ISPs), the data centers and Metropolitan Area Network (MAN).

▪ Initially, service providers at inner-edge only offer the fundamental

infrastructures for connecting local networks to the global Internet.

▪ However, the recent needs in improving the Quality of Experience (QoE) of Web
services have motivated the geo-distributed caching and processing mechanism
at the network data centers of WAN
 Hierarchy of Fog and Edge Computing
Middle-edge:

▪ Middle-edge corresponds to the environment which consists of two types of

networks—Local Area Network (LAN) and cellular network.

1. LAN includes ethernet, Wireless LAN (WLAN) and Campus Area Network (CAN).

2. Cellular network consists of the macrocell, microcell, picocell, and -femtocell.

✔ Macrocells typically cover a wide area, such as a city or suburban region.

✔ Microcells have a smaller coverage area

✔ Picocells have an even smaller coverage area than microcells – Indoor Location
 Hierarchy of Fog and Edge Computing
Outer-edge:

▪ Outer-edge, which is also known as extreme-edge, far-edge or mist

▪ It represents the front-end of the IoT network which consists of three types of
devices—constraint devices, integrated devices and IP gateway devices.

▪ Constraint devices such as sensors or actuators are usually operated by

microcontrollers,

▪ Integrated devices are the devices operated by the processors.

▪ IP gateway devices are also known as hubs, which act as the mediator between the
constrained devices and the middle-edge devices
Module I –L9
LO1: Define Edge Orchestration

LO2:Demonstrate resource allocation to edge devices

Resource Management and Edge
Orchestration
▪ Resource management and edge orchestration mechanisms are critical for
efficiently allocating and coordinating computing, storage, and networking
resources in edge computing environments.

▪ Allocation, optimization, and monitoring of computing, storage, and networking

resources within edge nodes or devices.

▪ Monitor resource health and performance for proactive maintenance.

▪ Ensure that adequate resources are available for critical tasks.

Resource Management and Edge
Orchestration
▪ Edge orchestration refers to the coordination and management of diverse
edge devices and resources to enable the execution of applications and
services.

▪ Automate deployment, scaling, and updates of edge applications.

▪ Enable centralized control and monitoring of edge infrastructure.

▪ Resource management techniques may include load balancing, resource

provisioning, virtualization to efficiently manage and allocate resources.
Cloud Vs Edge Vs Fog
Cloud Vs Edge Vs Fog
Cloud Vs Edge Vs Fog