Module
Module Name Hours Marks
No.
1 Introduction to Cloud Computing 5 14
2 Virtualization & Scalability 6 15
3 Cloud Services & Business Value in AI 8 20
4 Cloud Deployment & Edge AI 7 18
5 Security & Ethics 7 18
Architectures for Cloud-Based AI
6 6 15
Applications
Total Hours and Marks 39 100*
Module 1
Introduction to Cloud
Computing
SUBJECT: CLOUD COMPUTING
MR. AJINKYA VALANJOO
Overview
Introduction to Cloud Computing
Cloud characteristics
Cloud computing components
Comparing of cloud computing with Peer to peer architecture, Client server, Distributed, Grid
Cloud Deployment Model – Public, Private, Community, Hybrid
Service Models – SaaS, PaaS, IaaS
Need of cloud
Suppose you want to host a website, you need following things:
Need of cloud
Problems we were facing before cloud computing are:
why cloud
Introduction to cloud
What is Computing?
Includes designing and building hardware / software for a wide range of purposes, processing,
structuring and managing various kinds of information
Distributed Computing
Grid Computing
Cluster Computing
Utility Computing
Cloud Computing
Cloud computing
● Storting data/applications on
remote servers
● Processing data/applications
from servers
● Accessing data/applications
via internet
Introduction to cloud
● The term Cloud refers to a Network or Internet.
● A cloud is something that exists in a remote area.
● Cloud can provide services over public and private networks,
i.e., WAN, LAN or VPN.
● Applications such as e-mail, web conferencing, customer
relationship management (CRM) execute on cloud.
Types of Cloud
Cloud Deployment Models
❑ Public Cloud
❑ Private Cloud
❑ Community Cloud
❑ Hybrid Cloud
1. Public Cloud
● The public cloud infrastructure makes systems and
services available to anyone with an internet connection.
● Because of its openness, the public cloud is much less
secure.
Example: Microsoft, Google App Engine, Windows Azure
Services Platform.
Public Cloud
The cloud infrastructure is made available to the general
public or a large industry group and is owned by an
organization selling cloud services.
E.g. Microsoft Azure, AWS, Google clouds etc.
1. Public Cloud
2. Private cloud
● The private cloud infrastructure makes systems and services
available to anyone within an organization.
● Because of this private characteristic, it is more secure.
Examples: OpenNebula, OpenStack are some examples of this
style of deployment.
Private Cloud
The cloud infrastructure is operated solely for a single
organization. It may be managed by the organization or a
third party, and may exist on-premises or off-premises.
It has high level of security and privacy.
E.g. HP data centers
2. Private cloud
3. Hybrid Cloud
● Two or more clouds are combined to make up a hybrid
cloud infrastructure.
● In the majority of cases, public and private clouds blend
to make a hybrid cloud, where the private cloud handles
crucial tasks and the public cloud handles non-critical
tasks.
Examples: Google Application Suite (Gmail, Google Apps, and
Google Drive), Office 365 (MS Office on the Web and One
Drive), Amazon Web Services.
3. Hybrid Cloud
4. Community Cloud
● The community cloud infrastructure makes systems and services
available to anyone within a group of organizations.
● It is owned, managed, and operated by one or more
organizations in the community, a third party, or a combination
of them.
● It allows us to share cloud resources, infrastructure, and other
capabilities among various organizations.
Example: Health Care community cloud
4. Community Cloud
Community clouds are multi-tenant platforms and make it
possible for various companies to function on a shared
platform.
Community Cloud
The cloud infrastructure is made available to multiple
organizations where they share resources and services
based on common operational and regulatory
requirements.
E.g. IBM SoftLayer cloud
Cloud Computing
“The cloud computing service has grown
nearly 80% year-over-year in the last two
quarters, and is on pace to hit $7.8 billion in
revenue in 2015, four times the 2012 sales of
$1.8 billion.”
Cloud Computing
NITS (National Institute of Standards and Technology ) defines Cloud Computing as;
“ Cloud computing is a model for enabling ubiquitous, convenient,
on-demand network access to a shared pool of configurable
computing resources (e.g networks, servers, storage, applications,
and services) that can be rapidly provisioned and released with
minimal management effort or service provider interaction. ”
Characteristics
❑ On-demand self-service:
A consumer can unilaterally provision computing capabilities, such as server time and network
storage, as needed automatically without requiring human interaction with each service
provider
❑ Broad network access:
Capabilities are available over the network and accessed through standard mechanisms that
promote use by heterogeneous thin or thick client platforms (e.g., mobile phones, tablets,
laptops, and workstations).
❑ Resource pooling
The provider’s computing resources are pooled to serve multiple consumers using a multi-tenant
model, with different physical and virtual resources dynamically assigned and reassigned
according to consumer demand
Continue..
❑ Rapid elasticity:
- Capabilities can be rapidly and elastically provisioned - in some cases automatically - to quickly
scale out; and rapidly released to quickly scale in.
-To the consumer, the capabilities available for provisioning often appear to be unlimited and can
be purchased in any quantity at any time
❑ Measured service:
-Cloud systems automatically control and optimize resource usage by leveraging a metering
capability at some level of abstraction appropriate to the type of service
- Resource usage can be monitored, controlled, and reported - providing transparency for both
the provider and consumer of the service
Some other common characteristics
▪ Massive scale
▪Resilient computing
▪ Homogeneity
▪Geographic distribution
▪Virtualization
▪Service orientation
▪Low cost software
▪Advanced security
Cloud Services
Cloud services covers a wide range of services that a service
provider delivers to the customer via the internet.
Cloud Service models
Three main cloud services:
Software as a Platform as a Infrastructure
Service Service as a Service
Types of Service model
Types of Service model
Types of Service model
Types of Cloud Services
Types of Service model
1. IaaS (Infrastructure as a Service)
Examples:
AWS EC2, Microsoft Azure, Google Compute Engine
Who uses it:
IT Administrators
What you get:
•Virtual machines
•Storage
•Networking
Your responsibility:
•Operating system
•Runtime
•Middleware
•Applications
🔹 Most control, but more management effort
2. PaaS (Platform as a Service)
Examples:
Heroku, AWS Elastic Beanstalk, Google App Engine
Who uses it:
Software Developers
What you get:
•Infrastructure + OS
•Runtime environment
•Development tools
Your responsibility:
•Application code
•Data
🔹 Balanced control, faster development
3. SaaS (Software as a Service)
Examples:
Gmail, Trello, Salesforce CRM, EventPro, Office 365, Google Docs
Who uses it:
End Users
What you get:
•Fully managed software
•Accessible via browser
Your responsibility:
•Just use the application
🔹 Least control, maximum convenience
Key Concept Shown on the Right
Less control, more convenience
•Top (SaaS):
•Bottom (IaaS): More control, more responsibility
One-line summary (useful for exams/interviews):
IaaS gives infrastructure, PaaS gives a development platform, and SaaS gives ready-to-use software.
Cloud Service Models
1. Cloud Software as a Service (SaaS)
2. Cloud Platform as a Service (PaaS)
3. Cloud Infrastructure as a Service (IaaS)
Software as a Service (SaaS)
▪The capability provided to the consumer is to use the provider’s applications
running on a cloud infrastructure.
▪ The applications are accessible from various client devices through a thin client
interface such as a web browser (e.g., web-based email).
▪ The consumer does not manage or control the underlying cloud infrastructure
including network, servers, operating systems, storage, or even individual
application capabilities, with the possible exception of limited user-specific
application configuration settings.
Platform as a Service (PaaS)
▪The capability provided to the consumer is to deploy onto the cloud
infrastructure consumer created or acquired applications created using
programming languages and tools supported by the provider
▪ The consumer does not manage or control the underlying cloud infrastructure
including network, servers, operating systems, or storage, but has control over
the deployed applications and possibly application hosting environment
configurations
Infrastructure as a Service (IaaS)
▪The capability provided to the consumer is to provision processing, storage,
networks, and other fundamental computing resources.
▪ Consumer is able to deploy and run arbitrary 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, deployed applications, and
possibly limited control of select networking components (e.g., host firewalls)
Centralized Vs Distributed Computing
Distributed computing
▪ Controlling and managing is distributed
▪ Uni processor is not a concept
▪ Distributed Processor carries its own local memory and storage
▪ It provides computing services in distributed manner
▪ Each processor communicates with another processor through high speed lines
▪ Components: Workstations, Servers, Personal assistant devices
▪ Examples: Internet, ATM machines, Intranet
Distributed applications
1. Clients invoke individual servers
What’s happening in the image
•Clients (left side) send a request (invocation) directly to a specific server.
•The server processes the request.
•The result/response is sent back to the same client.
•Each client communicates independently with the server—there’s no intermediary like a load balancer or broker.
Key points explained
•🔹 Direct communication: Clients know the server’s address and call it directly.
•🔹 Synchronous interaction: Typically, the client waits for the server’s response.
•🔹 Tightly coupled: Clients are dependent on individual servers being available.
•🔹 Simple architecture: Easy to design and understand.
Advantages
•Simple to implement
•Low overhead
•Faster communication for small systems
Limitations
•Poor scalability when clients increase
•Server can become a bottleneck
•Failure of a server directly affects clients
Typical use cases
•Small-scale applications
•Early distributed systems
•Simple web or database client-server setups
2. A typical distributed application based on peer processes
What the diagram represents
•Each peer runs an application and can act as both client and server.
•Peers directly communicate with one another (no central server).
•Sharable objects/resources are distributed across peers.
•New peers (Peers 5 … N) can join dynamically and participate in sharing.
How it works
•A peer can request data/services from another peer.
•The same peer can serve data/services to others.
•Communication is many-to-many, unlike client–server (one-to-many).
Key characteristics
•🔹 Decentralized control – no single point of failure
•🔹 Scalable – adding peers increases resources
•🔹 Resource sharing – data, computation, or services
•🔹 Dynamic topology – peers can join/leave anytime
Advantages
•High scalability
•Fault tolerant (system continues if some peers fail)
•Better resource utilization
Limitations
•Complex management
•Security and trust issues
•Data consistency is harder to maintain
Examples
•BitTorrent
•Blockchain networks
•Distributed file sharing systems
Exam-oriented summary (2–3 lines)
In a peer-based distributed application, all nodes act as equals and can function as both clients and servers.
Resources and services are shared directly among peers without centralized coordination.
Grid Computing
▪ Grid computing harnesses unused processing cycles of all computers in a network for solving
problems too intensive for any stand-alone machine
▪
Application of Grid computing
▪Today’s Science/Research is based on computations, data analysis, data visualization &
collaborations
▪ Computer Simulations & Modelling are more cost effective than experimental methods
▪ Scientific and Engineering problems are becoming more complex & users need more accurate,
precise solutions to their problems in shortest possible time
▪ Data Visualization is becoming very important
▪ Exploiting under utilized resources
Cluster Computing
▪ A cluster is a type of parallel or distributed computer system, which consists of a collection of
inter-connected stand-alone computers working together as a single integrated computing resource
▪ Key components of a cluster include multiple standalone computers (PCs, Workstations, or SMPs),
operating systems, high-performance interconnects, middleware, parallel programming environments,
and applications.
▪Clusters are usually deployed to improve speed and/or reliability over that provided by a single
computer
▪Basic building blocks of clusters are broken down into multiple categories:
• Cluster Nodes
• Cluster Network
• Network Characterization
Utility Computing
▪ Utility Computing is purely a concept which cloud computing practically implements
▪ Utility computing is a service provisioning model in which a service provider makes computing
resources and infrastructure management available to the customer as needed, and charges them for
specific usage rather than a flat rate.
▪Some highlights:
a) Pay-for-use Pricing Business Model
b) Data Center Virtualization and Provisioning
c) Solves Resource Utilization Problem
d) Outsourcing
e) Web Services Delivery
f) Automation
▪Drawbacks: Data Backup , Data Security, Partner Competency, Defining SLA
Advantages
▪Lower Computer cost
▪Improved software updates
▪Improved document format capabilities
▪Unlimited storage capacity
▪Increased data reliability
▪Universal information access
▪Latest version availability
▪Easier group collaboration
▪Device independance
Drawback
▪Requires constant internet connection
▪Does not work well with low speed
▪Features might be limited
▪Can be slow
▪Security
▪Stored data can be lost
Virtualization
Introduction to Virtualization
● Virtualization is the process of creating a
virtual environment of something which
may include hardware platforms, storage
devices, OS, network resources, etc.
● It is the process of creating a virtual instance of a device, such
as a virtual server or virtual operating system.
Need of Virtualization
Virtualization
● Virtualization creates virtual layer using hypervisor software
which manages resources assigned to virtual instances.
● The newly formed virtual representation is called as virtual
machine (VM).
● Virtual machine is an emulation or virtual representation of
physical system.
Benefits of Virtualization
1. More flexible and efficient allocation of resources.
2. Enhance development productivity.
3. It lowers the cost of IT infrastructure.
4. Remote access and rapid scalability.
5. High availability and disaster recovery.
6. Enables running multiple operating systems.
Hypervisor
● The virtual machine is referred as a guest
machine and physical machine on which
the virtual machine is created is known
as host machine.
● The virtual machine is managed by a
software or firmware, which is known as
Hypervisor.
Hypervisor
Hypervisor is a firmware
or low-level program that
acts as a Virtual Machine
Manager (VMM).
Hypervisor
There are two types of hypervisor:
[Link] 1 hypervisor / Native/Bare-Metal Hypervisors
[Link] 2 hypervisor / Hosted Hypervisors
Hypervisor
1. Type 1 Hypervisor
● They are deployed directly over the host hardware.
● Direct access to the hardware without any underlying OS
or device drivers makes such hypervisors highly efficient
for enterprise computing.
● VMware ESXi, Microsoft Hyper-V, Oracle VM, and Xen
are examples of type 1 hypervisors.
2. Type 2 Hypervisor
● They run as an application over a traditional OS.
● Developers, security professionals, or users who need to
access applications only available on select OS versions
often rely on type 2 hypervisors for their operations.
● KVM, VMware Server and Workstation, Microsoft Virtual
PC, Oracle VM VirtualBox, and QEMU are popular type 2
hypervisors.
Types of virtualization
1. Desktop virtualization
● It allows us to run multiple desktop OS in each VM.
● t allows the user to access their desktop virtually, from any
location by a different machine.
● Users who want specific operating systems other than Windows
Server will need to have a virtual desktop.
● Main benefits of desktop
virtualization are user mobility,
portability, easy management of
software installation, updates, and
patches.
2. Network virtualization
● It basically refers to a system where one can run multiple
virtual networks at the same time and each of these
networks have a separate control system and data plan.
● It also helps you keep a better track of things on the
network and allow for good supervision as well as
identification of the data usage.
● Example : firewall
3. Storage virtualization
● Storage virtualization is an array of servers that are managed
by a virtual storage system.
● The servers aren’t aware of exactly where their data is stored,
and instead function more like worker bees in a hive.
4. Application virtualization
● Application virtualization helps a user to
have remote access of an application from
a server.
● The server stores all personal information
and other characteristics of the application
but can still run on a local workstation
through the internet.
● Example of this would be a user who
needs to run two different versions of the
same software.
Major Cloud Providers – Overview
AWS, Microsoft Azure, and Google Cloud Platform (GCP)
Cloud computing services are dominated by three major global providers:
Amazon Web Services (AWS), Microsoft Azure, and Google Cloud
Platform (GCP). These providers offer scalable, on-demand infrastructure
and platform services used by startups, enterprises, and governments
worldwide.
Introduction
•Cloud computing is dominated by a few major global providers
•These providers offer:
• On-demand computing resources
• Scalable infrastructure
• Pay-as-you-use pricing
•The top three cloud providers are:
• Amazon Web Services (AWS)
• Microsoft Azure
• Google Cloud Platform (GCP)
Amazon Web Services (AWS) – Overview
•Launched in 2006
•Provided by Amazon
•World’s largest cloud service provider
•Offers the widest range of cloud services
•Used by startups, enterprises, and governments
AWS – Key Services
•Compute: EC2, Lambda
•Storage: S3, EBS, Glacier
•Database: RDS, DynamoDB
•Networking: VPC, CloudFront
•Security: IAM, Shield
Strength: Mature platform with global reach
AWS – Use Cases
•Web and mobile applications
•Big data processing
•Machine learning and AI
•Backup and disaster recovery
•Startups and large-scale enterprises
Microsoft Azure – Overview
•Launched in 2010
•Provided by Microsoft
•Strong focus on enterprise customers
•Excellent support for hybrid cloud
•Seamless integration with Microsoft tools
Azure – Key Services
•Compute: Azure Virtual Machines, Azure Functions
•Storage: Blob Storage, Disk Storage
•Database: Azure SQL Database, Cosmos DB
•Networking: Virtual Network, Load Balancer
•Security: Azure Active Directory
Strength: Best for Microsoft-based environments
Azure – Use Cases
•Enterprise applications
•Hybrid cloud deployments
•Windows and .NET workloads
•Government and regulated industries
•Corporate IT systems
Google Cloud Platform (GCP) – Overview
•Launched in 2011
•Provided by Google
•Built on Google’s global infrastructure
•Known for data analytics and AI
•Fast-growing cloud provider
GCP – Key Services
•Compute: Compute Engine, Cloud Functions
•Storage: Google Cloud Storage
•Database: BigQuery, Cloud SQL
•AI/ML: TensorFlow, Vertex AI
•Networking: VPC, Cloud Load Balancing
Strength: Data analytics and machine learning
GCP – Use Cases
•Big data analytics
•AI and machine learning applications
•Cloud-native development
•High-performance computing
•Research and data-driven workloads
Comparison of Cloud Providers
Feature AWS Azure GCP
Market Share Highest Second Growing
Best Known For Service variety Enterprise integration Data & AI
Hybrid Cloud Limited Strong Moderate
Popular Service EC2, S3 Azure VM, SQL BigQuery
Advantages Summary
•AWS: Most comprehensive and reliable
•Azure: Best for enterprise and hybrid cloud
•GCP: Best for analytics, AI, and ML
Conclusion
•All three providers offer powerful cloud services
•Choice depends on:
• Business requirements
• Existing technologies
• Cost and performance needs
•Understanding these platforms is essential for cloud
professionals
The Role of Cloud in AI
Introduction
•Artificial Intelligence requires:
• High computational power
• Large data storage
• Fast and parallel processing
•Traditional on-premise systems are:
• Costly
• Difficult to scale
•Cloud computing provides an ideal platform for AI workloads
Why Cloud is Important for AI
•AI workloads are:
• Compute intensive
• Data intensive
•Cloud provides:
• Elastic infrastructure
• On-demand resources
• Pay-as-you-use model
•Enables organizations to adopt AI without heavy capital investment
Enabling Scalable Training of AI Models
AI model training requires:
•Large datasets
•GPUs / TPUs
•Parallel processing
Cloud enables:
•Horizontal and vertical scaling
•Distributed training
•Faster model convergence
Example: Training deep learning models using thousands of GPUs for a short
duration.
1. GPU (Graphics Processing Unit)
Definition
A GPU is a specialized processor designed to perform parallel computations efficiently.
In AI, GPUs accelerate training and inference of machine learning and deep learning models.
Why GPUs are Used in AI
•AI involves matrix and vector operations
•GPUs can execute thousands of operations simultaneously
•Much faster than CPUs for deep learning workloads
Key Features
•Massive parallelism
•High memory bandwidth
•Suitable for:
• Neural network training
• Image and video processing
• Scientific simulations
Cloud Examples
•AWS: EC2 GPU instances
•Azure: NV-series, NC-series VMs
•GCP: NVIDIA GPU-enabled Compute Engine
Training Data
|
v
+------------+
| Cloud GPU |
| Instance |
+------------+
|
v
Trained AI Model
2. TPU (Tensor Processing Unit)
Definition
A TPU is a custom-built AI accelerator developed by Google, specifically optimized for tensor
operations used in deep learning.
Why TPUs are Used
•Designed for deep learning workloads
•Extremely efficient for:
• Neural networks
• TensorFlow models
Key Features
•High performance per watt
•Optimized for matrix multiplication
•Faster training and inference than GPUs for supported workloads
Cloud Availability
•Google Cloud Platform (GCP) provides TPUs
•Integrated with TensorFlow and Vertex A
Training Data
|
v
+------------+
| Cloud TPU |
| Accelerator|
+------------+
|
v
Trained AI Model
Diagram – Scalable AI Training on Cloud
Training Data
|
v
+------------------+
| Cloud Storage |
| (S3 / Blob / GCS)|
+------------------+
|
v
+-------------------------+
| GPU / TPU Compute Nodes |
| (VMs / Containers) |
+-------------------------+
|
v
Trained AI Model
Cloud-Based Model Hosting
After training, AI models must be:
•Deployed
•Scalable
•Highly available
Cloud provides:
•Managed model hosting
•Auto-scaling
•Load balancing
•API-based access
Diagram – AI Model Hosting
Users / Applications
|
v
+----------------+
| Load Balancer |
+----------------+
|
v
+----------------------+
| Cloud AI Model |
| (Hosted as Service) |
+----------------------+
|
v
Predictions
Data Processing for AI
AI systems depend on massive data processing:
•Data collection
•Data cleaning
•Feature extraction
•Analysis
Cloud enables:
Distributed data processing
Big Data frameworks
Real-time and batch analytics
Diagram – Cloud-Based Data Processing
Data Sources
(IoT, Logs, Apps)
| A Cloud Data Lake is a centralized cloud-based storage repository that
v allows organizations to store structured, semi-structured, and
+-------------------+ unstructured data at any scale, in raw format, for analytics, big data
| Cloud Data Lake | processing, and AI/ML applications.
+-------------------+
|
Data Processing (Spark / Hadoop)
v
+-------------------+
| Data Processing | What is Data Processing in Cloud?
| (Spark / Hadoop) | The Data Processing layer in cloud computing is responsible for cleaning,
+-------------------+ transforming, and analyzing large volumes of data stored in a Cloud Data Lake
| before it is used for analytics or AI model training.
v
AI Training
Comparison – Cloud vs Grid Computing
Feature Cloud Computing Grid Computing
Resource Access On-demand Shared
Scalability Highly elastic Limited
Cost Model Pay-per-use Fixed / shared
Management Easy Complex
AI Suitability High Moderate
Cloud vs Grid – Key Differences
•Cloud:
• Service-oriented
• Supports virtualization and containers
• Best suited for AI workloads
•Grid:
• Resource-sharing oriented
• Limited elasticity
• Less flexible for AI training
Business Drivers for AI Adoption Using Cloud
Key Business Drivers:
•Cost Reduction
• No upfront infrastructure cost
•Faster Innovation
• Rapid experimentation
•Scalability
• Easy handling of growing data and users
•Access to Advanced AI Services
• Pre-built ML and AI APIs
AI Use Cases Enabled by Cloud
Major Use Cases:
•Healthcare
• Medical image analysis
• Disease prediction
•Finance
• Fraud detection
• Credit risk analysis
•Retail
• Recommendation systems
• Demand forecasting
•Smart Cities
• Traffic prediction
• Energy optimization
Diagram – AI Use Cases via Cloud
Cloud AI Platform
(Compute + Storage + AI)
|
--------------------------------
| | | |
Healthcare Finance Retail Smart Cities
Advantages of Cloud-Based AI
•High scalability
•Reduced deployment time
•Global accessibility
•High reliability
•Better performance for AI workloads
Conclusion
•Cloud computing acts as the foundation for modern AI systems
•Enables:
• Scalable training
• Efficient model hosting
• Large-scale data processing
•Cloud is more suitable than grid computing for AI workloads
•Drives AI adoption across industries
Thank You!