BCS601 | CLOUD COMPUTING |

Module – 03

Cloud Platform Architecture over Virtualized Datacenters

Cloud Computing and Service Models

Definition and Characteristics of Cloud Computing

Cloud computing is a technology that delivers computing services such as storage, servers,
databases, networking, software, and analytics over the internet (the cloud) instead of local
servers or personal devices.

It enables on-demand access to resources with minimal management effort.

Characteristics of Cloud Computing:

 On-Demand Self-Service: Users can provision resources automatically without human

intervention.
 Broad Network Access: Services are accessible from anywhere via the internet.
 Resource Pooling: Cloud providers use multi-tenancy to serve multiple customers with
shared computing resources.
 Rapid Elasticity: Resources can scale up or down dynamically based on demand.
 Measured Service: Usage is monitored and billed based on consumption.

Evolution from Traditional Datacenters to Cloud

Traditional datacenters required organizations to own and maintain physical servers,

networking hardware, and software.

These setups were costly, lacked scalability, and required significant IT expertise for
management.

Page 1
 BCS601 | CLOUD COMPUTING |

Transition to Cloud Computing:

 Virtualization enabled more efficient resource utilization.

 Internet bandwidth improvements allowed cloud-based applications to run
smoothly.
 Pay-as-you-go models reduced capital expenditure (CAPEX) and increased
operational efficiency.

Cloud computing evolved from these limitations, allowing businesses to focus on innovation
rather than IT infrastructure management.

Virtualization and its Role in Cloud Computing

Virtualization is the technology that enables cloud computing by abstracting hardware resources.
It allows multiple virtual machines (VMs) to run on a single physical server, improving
efficiency and flexibility.

Types of Virtualization:

 Server Virtualization: Divides a physical server into multiple VMs.

 Storage Virtualization: Combines multiple storage devices into a single storage pool.
 Network Virtualization: Creates virtual network layers for better traffic
management.

Benefits of Virtualization in Cloud Computing:

 Cost Efficiency: Reduces hardware dependency.

 Scalability: Allows dynamic allocation of resources.
 Flexibility: Supports multiple OS environments on the same hardware.

Page 2
 BCS601 | CLOUD COMPUTING |

Cloud Service Models (IaaS, PaaS, SaaS)

Cloud services are categorized into three main models, each serving different user needs:

1. Infrastructure as a Service (IaaS):

 Provides virtualized computing resources such as servers, storage, and

networking.
 Example: Amazon Web Services (AWS) EC2, Google Compute Engine (GCE),
Microsoft Azure.

2. Platform as a Service (PaaS):

 Offers a development environment with pre-configured tools, databases, and

middleware.

Page 3
 BCS601 | CLOUD COMPUTING |

 Example: Google App Engine, Microsoft Azure App Services, AWS Elastic
Beanstalk.

3. Software as a Service (SaaS):

 Delivers fully managed software applications over the internet.

 Example: Gmail, Dropbox, Microsoft 365, Salesforce.

Cloud Deployment Models (Public, Private, Hybrid, Multi-Cloud)

Cloud environments are deployed based on organizational needs and security requirements:

1. Public Cloud:

 Hosted and managed by third-party cloud providers.

 Example: AWS, Google Cloud, Microsoft Azure.
 Benefits: Cost-effective, scalable, easy to deploy.

2. Private Cloud:

 Dedicated cloud environment for a single organization.

 Example: VMware vSphere, OpenStack.
 Benefits: High security, better control, compliance-friendly.

Page 4
 BCS601 | CLOUD COMPUTING |

3. Hybrid Cloud:

 Combines public and private cloud services.

 Benefits: Balances security with scalability.

4. Multi-Cloud:

 Uses multiple cloud providers for different services.

 Benefits: Avoids vendor lock-in, improves resilience.

Cloud Computing and Service Models

Cloud computing is categorized into three primary service models: Infrastructure as a Service
(IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). Each model serves
different purposes, catering to various business and IT needs.

Page 5
 BCS601 | CLOUD COMPUTING |

Infrastructure as a Service (IaaS)

IaaS provides virtualized computing resources over the internet, offering businesses flexible
and scalable infrastructure without the need for on-premises hardware.

Components of IaaS:

1. Compute: Virtual Machines (VMs) and containers for running applications (e.g., AWS
EC2, Google Compute Engine).
2. Storage: Block, object, and file storage for data management (e.g., Amazon S3, Google
Cloud Storage).
3. Networking: Virtual networks, firewalls, load balancers (e.g., AWS VPC, Azure
Virtual Network).

Examples of IaaS Providers:

 Amazon Web Services (AWS) – EC2, S3, VPC

 Microsoft Azure – Virtual Machines, Blob Storage
 Google Cloud – Compute Engine, Cloud Storage Use

Cases of IaaS:

 Hosting websites and applications

 High-performance computing (HPC)
 Disaster recovery and backup solutions

Platform as a Service (PaaS)

PaaS provides a development and deployment environment with pre-configured tools,

databases, and frameworks, reducing the complexity of software development.

Features of PaaS:

Page 6
 BCS601 | CLOUD COMPUTING |

1. Development Tools: SDKs, APIs, code repositories (e.g., AWS Lambda, Google App
Engine).
2. Database Management: Managed databases for application storage (e.g.,
Firebase, AWS RDS).
3. Application Hosting: Middleware and runtime environments (e.g., Azure App
Services).

Examples of PaaS Providers:

 Google App Engine

 AWS Elastic Beanstalk
 Microsoft Azure App Services

Use Cases of PaaS:

 Rapid application development

 Microservices architecture
 Continuous Integration/Continuous Deployment (CI/CD)

Software as a Service (SaaS)

SaaS provides fully managed software applications over the internet, eliminating the need for
users to install or maintain software locally.

Characteristics of SaaS:

1. Subscription-Based: Pay-as-you-go model for flexibility.

2. Accessible Anywhere: Users can access applications via web browsers.
3. Automatic Updates: No need for manual software upgrades.

Examples of SaaS Applications:

 Google Workspace (Gmail, Google Drive, Docs)

Page 7
 BCS601 | CLOUD COMPUTING |

 Microsoft 365 (Word, Excel, Teams)

 Salesforce (CRM software)

Use Cases of SaaS:

 Cloud-based email and collaboration tools

 Customer relationship management (CRM)
 Online productivity suites

Benefits and Challenges of Each Service Model

Case Studies of Service Models

Case Study 1: Netflix (IaaS)

Netflix runs its streaming service on Amazon Web Services (AWS) using EC2 for compute and
S3 for storage. This allows Netflix to handle high traffic and scale dynamically.

Case Study 2: Heroku (PaaS)

Heroku, a cloud platform for developers, provides a fully managed runtime environment that
simplifies application deployment, allowing startups and enterprises to quickly build and scale
applications.

Page 8
 BCS601 | CLOUD COMPUTING |

Case Study 3: Google Workspace (SaaS)

Google Workspace (formerly G Suite) offers cloud-based email, document editing, and
collaboration tools, reducing IT overhead and improving productivity.

Data Center Design and Interconnection Networks

Cloud data centers form the backbone of cloud computing, offering scalable, secure, and efficient
computing environments.

The design and interconnection of these data centers impact performance, resource utilization,
and energy efficiency.

Modern Cloud Data Center Architecture

A modern cloud data center is built on a highly scalable and modular architecture that supports
virtualization, high-speed networking, and distributed computing.

Page 9
 BCS601 | CLOUD COMPUTING |

Components of a Cloud Data Center:

 Compute Infrastructure: Virtualized servers that host applications and services.

 Storage Infrastructure: Distributed storage systems (e.g., SAN, NAS, object
storage).
 Networking Infrastructure: High-speed networks with software-defined
capabilities.
 Security & Management: Firewalls, intrusion detection, and monitoring tools.

Characteristics of Cloud Data Centers:

 Scalability: Ability to add or remove resources dynamically.

 High Availability: Redundant components to ensure minimal downtime.
 Energy Efficiency: Optimized power and cooling mechanisms.

Page 10
 BCS601 | CLOUD COMPUTING |

Server Virtualization and Resource Pooling

Server virtualization allows multiple virtual machines (VMs) to run on a single physical server,
improving resource utilization and reducing costs.

Virtualization Technologies:

 Hypervisors: Software that manages VMs (e.g., VMware ESXi, KVM, Microsoft
Hyper-V).
 Containers: Lightweight alternatives to VMs (e.g., Docker, Kubernetes).
 Resource Pooling: Dynamically allocating computing, storage, and network
resources across multiple users.

Benefits of Virtualization:

 Improved hardware utilization through VM consolidation.

 Greater flexibility and scalability in resource management.
 Reduced operational costs due to better resource allocation.

Software-Defined Networking (SDN) and Network Virtualization

Traditional data center networks are hardware-driven and static, making them complex to
manage. SDN and network virtualization address these limitations.

Software-Defined Networking (SDN):

 Decouples network control from hardware using a centralized controller.

 Enables programmable and dynamic network configuration.
 Example: OpenFlow protocol, Cisco ACI, VMware NSX.

Network Virtualization:

 Creates logical network segments on shared physical infrastructure.

 Supports multi-tenant environments by isolating traffic.

Page 11
 BCS601 | CLOUD COMPUTING |

 Examples: VXLAN, NVGRE, MPLS.

Benefits of SDN & Network Virtualization:

 Improved network flexibility and automation.

 Better security and traffic management.
 Scalability to handle dynamic cloud workloads.

Data Center Network (DCN) Topologies

DCN topologies define how servers, switches, and storage devices are interconnected for
scalability and performance.

Common DCN Architectures:

These architectures enhance data transfer efficiency, fault tolerance, and network scalability.

Storage Area Networks (SAN) and Network Attached Storage (NAS)

Storage systems in cloud data centers must be highly available, scalable, and efficient. Storage

Solutions in Data Centers:

Page 12
 BCS601 | CLOUD COMPUTING |

Differences:

 SAN is block-level storage, requiring high-speed connectivity (e.g., Fibre Channel,

iSCSI).
 NAS is file-based storage, accessible over standard networks (e.g., NFS, SMB).

Energy Efficiency and Green Computing in Data Centers

With the growing demand for cloud computing, energy efficiency has become a critical aspect of
modern data centers.

Green Computing Strategies:

1. Energy-Efficient Hardware: Low-power processors, SSDs, and smart cooling

systems.
2. Virtualization & Resource Optimization: Reducing the number of physical servers
through consolidation.
3. Renewable Energy Integration: Using solar, wind, and hydroelectric power.
4. Advanced Cooling Techniques: Liquid cooling, free air cooling, and AI-driven
temperature management.

Benefits of Energy-Efficient Data Centers:

 Reduces operational costs and carbon footprint.

 Enhances sustainability and compliance with environmental regulations.
 Improves overall system reliability and longevity.

Page 13
 BCS601 | CLOUD COMPUTING |

Examples: Google and Microsoft operate carbon-neutral data centers powered by renewable
energy.

Architectural Design of Compute and Storage Clouds

Cloud computing architecture consists of compute and storage clouds, which provide
scalable, flexible, and high-performance infrastructure.

This chapter explores key architectural components, including virtual machines vs. containers,
serverless computing, distributed computing, and cloud-native storage solutions.

Page 14
 BCS601 | CLOUD COMPUTING |

Compute Clouds: Virtual Machines vs. Containers

Compute clouds provide on-demand processing power for running applications. Two primary
approaches dominate modern compute architectures: virtual machines (VMs) and containers.

Virtual Machines (VMs)

Definition: A VM is a software-based simulation of a physical computer, running an operating

system (OS) on a hypervisor.

Hypervisors: VMware ESXi, Microsoft Hyper-V, KVM, Xen.

Features:

o Full OS virtualization
o Hardware emulation and abstraction
o Strong isolation between workloads

Containers

Definition: A container is a lightweight, isolated application runtime environment that shares

the host OS kernel.

Popular Container Platforms: Docker, Kubernetes, OpenShift.

Features:

o Faster startup time compared to VMs

o Less overhead, as they share the same OS kernel
o Ideal for microservices architecture

Page 15
 BCS601 | CLOUD COMPUTING |

Comparison: VMs vs. Containers

Trend: Many enterprises combine VMs and containers for hybrid architectures, using tools like
Kubernetes for container orchestration.

Serverless Computing and Function-as-a-Service (FaaS) What is

Serverless Computing?

 Serverless computing allows developers to run code without managing

infrastructure.
 The cloud provider handles server provisioning, scaling, and maintenance.
 Example services: AWS Lambda, Google Cloud Functions, Azure Functions.

Function-as-a-Service (FaaS)

 A type of serverless computing where applications are broken into functions that run
on-demand.
 Functions execute only when triggered and scale automatically.

Advantages of Serverless Computing

✔ Auto-scaling: No need to manually manage servers.

✔ Cost-efficient: Pay only for execution time.

✔ Simplified deployment: No infrastructure management required.

Page 16
 BCS601 | CLOUD COMPUTING |

Use Cases of Serverless Computing

 Event-driven applications (e.g., processing IoT data).

 API backends (e.g., RESTful services).
 Real-time file processing (e.g., automatic image resizing).

Distributed and Parallel Computing in Cloud

Cloud computing enables distributed and parallel processing to handle large-scale computations
efficiently.

Concepts

 Distributed Computing: Multiple machines work together to solve a problem.

 Parallel Computing: Multiple processors execute tasks simultaneously.

Popular Distributed Computing Frameworks

Use Cases of Distributed Computing

 Big Data Analytics (e.g., processing large datasets with Hadoop).

 AI & Machine Learning (e.g., training deep learning models).
 Financial Modeling (e.g., Monte Carlo simulations).

Page 17
 BCS601 | CLOUD COMPUTING |

Storage Clouds: Object Storage, Block Storage, and File Storage

Cloud storage services provide highly scalable and durable storage solutions for different
workloads.

Types of Cloud Storage

Comparison of Storage Models

 Object storage is ideal for unstructured data (e.g., images, videos).

 Block storage is best for low-latency, high-performance applications.
 File storage is used for collaborative file-sharing environments.

Cloud-Native Storage Solutions

Cloud-native storage solutions are optimized for scalability, durability, and accessibility. Cloud

Storage Services

1. AWS S3 (Simple Storage Service)

o Object storage for scalable and secure data management.
o Supports versioning, replication, and encryption.
o Commonly used for big data analytics, backups, and machine learning
datasets.
2. Google Cloud Storage
o Multi-regional, highly available object storage.

Page 18
 BCS601 | CLOUD COMPUTING |

o Integrated with Google AI and BigQuery for analytics.

o Supports lifecycle management and fine-grained access controls.
3. Azure Blob Storage
o Designed for massive-scale object storage.
o Optimized for media streaming, data lakes, and disaster recovery.
o Supports tiered storage options (hot, cool, archive).

Benefits of Cloud-Native Storage

✔ Highly scalable: Can handle petabytes of data.

✔ Cost-efficient: Different storage classes optimize costs.
✔ Durable and secure: 99% durability.

Public Cloud Platforms: GAE, AWS, and Azure

Public cloud platforms provide on-demand computing resources and services to organizations
and developers worldwide.

This chapter explores the three major cloud providers—Google App Engine (GAE),
Amazon Web Services (AWS), and Microsoft Azure—their key services, and pricing
strategies.

Overview of Public Cloud Providers

Public cloud platforms offer scalability, security, and flexibility by delivering

Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS), and Software-as-a-
Service (SaaS) solutions.

Characteristics of Public Cloud Providers

✔ Pay-as-you-go model – Users only pay for what they use.

✔ Global data centers – Cloud providers have regions and availability zones to ensure

Page 19
 BCS601 | CLOUD COMPUTING |

reliability.
✔ Security and compliance – Built-in security, encryption, and compliance with GDPR, HIPAA,
and ISO standards.
✔ Automation and AI integration – Tools for auto-scaling, monitoring, and machine learning
integration.

The three major public cloud providers are:

1. Google Cloud Platform (GCP) – Known for AI/ML and data analytics.
2. Amazon Web Services (AWS) – The most extensive cloud platform with a vast
ecosystem.
3. Microsoft Azure – A strong enterprise cloud with integration into Microsoft
products.

Google App Engine (GAE)

Google App Engine (GAE) is a Platform-as-a-Service (PaaS) offering that allows developers
to build and deploy scalable applications without managing infrastructure.

Features of Google App Engine

 Supports multiple programming languages (Python, Java, Node.js, Go).

 Automatic scaling and load balancing.
 Built-in security and integration with other Google Cloud services.
 Managed databases like Google Cloud Datastore and Firebase.
 Zero server management – Google manages the underlying infrastructure.

Use Cases of Google App Engine

✔ Web applications – Hosting scalable web apps and APIs.

✔ Mobile backends – Powering mobile applications with serverless backends.
✔ Data processing and analytics – Running AI and ML-based apps.

Page 20
 BCS601 | CLOUD COMPUTING |

Deployment on Google App Engine

1. Write application code in a supported language.

2. Define configuration in app.yaml.
3. Deploy using gcloud app deploy.
4. Manage versions, traffic splitting, and monitoring via the Google Cloud Console.

Amazon Web Services (AWS)

Amazon Web Services (AWS) is the most widely used public cloud provider, offering a broad
range of services across compute, storage, databases, networking, AI, and analytics.

AWS Services

AWS Strengths

✔ Extensive service portfolio – Covers nearly every cloud use case.

✔ Largest global infrastructure – 30+ regions and 90+ availability zones.

✔ Strong ecosystem – Works with third-party tools, DevOps, AI/ML frameworks.

Page 21
 BCS601 | CLOUD COMPUTING |

Microsoft Azure

Microsoft Azure provides a broad set of cloud services focused on enterprise applications, AI,
and hybrid cloud solutions.

Azure Services

Azure Strengths

✔ Deep integration with Microsoft products – Works with Windows Server, Active Directory,
and Office 365.
✔ Hybrid cloud focus – Azure Arc for hybrid and multi-cloud environments.
✔ Strong enterprise adoption – Trusted by large corporations and government organizations.

Page 22
 BCS601 | CLOUD COMPUTING |

Pricing Models and Cost Optimization Strategies Cloud

Pricing Models

Cost Optimization Strategies

✔ Use Auto-Scaling – Scale resources dynamically based on demand.

✔ Leverage Spot Instances – Reduce compute costs for non-critical workloads.

✔ Optimize Storage – Use lifecycle policies to move cold data to cheaper tiers.
✔ Monitor Costs – Use AWS Cost Explorer, Azure Cost Management, Google Cloud Pricing
Calculator.
✔ Implement Infrastructure-as-Code (IaC) – Use Terraform, AWS CloudFormation to
manage resources efficiently.

Inter-Cloud Resource Management

Page 23
 BCS601 | CLOUD COMPUTING |

Inter-cloud resource management refers to the techniques and strategies used to coordinate,
allocate, and optimize resources across multiple cloud environments.

This chapter explores cloud federation, multi-cloud strategies, resource orchestration,

interoperability standards, Kubernetes-based container orchestration, and real-world case
studies on cross-cloud deployment.

Cloud Federation and Multi-Cloud Strategies

Cloud Federation

 Cloud federation is a model where multiple cloud providers or data centers

collaborate to share resources and workloads.
 Enables scalability, redundancy, and cost optimization.
 Common in government, research, and enterprise IT where resources are pooled for
efficiency.

Multi-Cloud Strategy

A multi-cloud strategy involves using multiple cloud providers (AWS, Azure, GCP) to avoid
vendor lock-in and improve availability, performance, and compliance.

Benefits of Multi-Cloud Strategies

Page 24
 BCS601 | CLOUD COMPUTING |

✔ Avoids vendor lock-in – Organizations are not dependent on a single provider.

✔ Improves redundancy – Ensures higher availability and failover capabilities.

✔ Optimizes costs – Leverages cost differences between cloud providers.

✔ Enhances performance – Deploy workloads in geographically closer data centers for lower
latency.

Resource Allocation and Orchestration across Multiple Clouds

Challenges in Multi-Cloud Resource Management

 Different APIs and configurations across providers.

 Data transfer costs and latency when moving workloads.
 Security and compliance across different cloud environments.
 Auto-scaling and load balancing across distributed clouds.

Techniques for Efficient Resource Allocation

Page 25
 BCS601 | CLOUD COMPUTING |

Cloud Interoperability Standards and Protocols

Cloud interoperability ensures that services from different cloud providers can work together
seamlessly.

Interoperability Challenges

Proprietary APIs – Each provider has different APIs, configurations, and services. Data
Portability – Moving workloads and data between clouds can be costly.
Networking and Security – Ensuring consistent access control, encryption, and firewall
policies.

Cloud Standards and Protocols for Interoperability

Page 26
 BCS601 | CLOUD COMPUTING |

Kubernetes and Container Orchestration for Multi-Cloud

Kubernetes is the leading container orchestration platform that enables automated

deployment, scaling, and management of containerized applications across multiple clouds.

Why Use Kubernetes for Multi-Cloud?

✔ Unified orchestration – Manage workloads across AWS, Azure, and GCP.

✔ Portability – Containers can be moved easily between cloud providers.

✔ Auto-scaling – Kubernetes adjusts resources dynamically.

✔ Service discovery and load balancing – Efficient traffic management across clouds. Multi-

Cloud Kubernetes Solutions

Example Multi-Cloud Deployment with Kubernetes

1. Deploy Kubernetes Clusters on AWS, Azure, and GCP.

2. Use KubeFed to federate clusters.
3. Configure service mesh (Istio) for secure inter-cloud communication.
4. Implement auto-scaling to optimize performance and cost.

Page 27
 BCS601 | CLOUD COMPUTING |

Case Studies on Cross-Cloud Deployment and Management Case

Study 1: Multi-Cloud Deployment with Kubernetes

 Company: Global e-commerce platform.

 Challenge: High traffic required failover across AWS, GCP, and Azure.
 Solution:
o Used Kubernetes and Istio for multi-cloud traffic routing.
o Implemented auto-scaling based on regional demand.
o Used Terraform for consistent deployment.
 Outcome: Improved availability, reduced latency, and optimized costs.

Case Study 2: Hybrid Cloud for Financial Services

 Company: A multinational bank.

 Challenge: Compliance required on-prem data storage while leveraging cloud
computing power.
 Solution:
o Used Google Anthos for hybrid cloud management.
o Hosted data-sensitive applications on-prem, with AI analytics in GCP.
o Implemented Azure Arc to manage workloads.
 Outcome: Achieved regulatory compliance while benefiting from cloud scalability.

Case Study 3: Cloud Bursting for AI Workloads

 Company: AI startup.
 Challenge: On-prem GPU clusters were expensive and had limited scalability.
 Solution:
o Used AWS EC2 Spot Instances for AI model training.
o Offloaded inference workloads to Google Cloud TPU for faster processing.
o Implemented autoscaling for cost optimization.
 Outcome: Reduced infrastructure costs by 40% while maintaining high
performance.

Page 28