Cloud Computing 601-Module 3

1. Private, Public, and Hybrid cloud deployment models

Cloud deployment models define how cloud services are made available to users. They are
mainly categorized into:
• Public Cloud
• Private Cloud
• Hybrid Cloud
Each model differs in terms of ownership, access level, cost, security, and control.

Public Cloud
• Definition:
Public cloud is a cloud environment open to general public and hosted over
the Internet by third-party providers.
• Ownership:
Owned and managed by external providers like AWS, Google Cloud, Microsoft Azure.
• Access:
Any user can subscribe and access the services via pay-per-use model.
• Advantages:
o Scalable on-demand
o Cost-effective for startups
o Easy to use and deploy globally
• Examples:
o Amazon Web Services (AWS)
o Google Cloud Platform (GCP)
o Microsoft Azure
o IBM Blue Cloud

Private Cloud
• Definition:
A private cloud is used exclusively by a single organization, often managed internally or
by a third party, and hosted within the organization’s firewall.
• Ownership:
Owned and controlled by the organization.
• Access:
Limited to internal users (employees, departments, etc.)

1
PREPARED BY: DR. JUNID
 • Advantages:
o Enhanced security and privacy
o Greater customization and control
o Complies with regulatory standards
• Examples:
o Government or banking sector using internal private infrastructure
o IBM’s Research Compute Cloud (RC2)

Hybrid Cloud
• Definition:
Hybrid cloud is a combination of public and private clouds that allows data and
applications to be shared between them.
• Ownership:
Managed by both organization and public provider.
• Access:
Internal + external access depending on configuration.
• Advantages:
o Flexibility: Critical workloads on private, non-sensitive on public
o Cost optimization
o Useful for bursting workloads into public cloud during peak demand
• Examples:
o A company using private cloud for sensitive data and public cloud for customer-
facing apps
o IBM RC2 (hybrid of 8 global research centers)

2
PREPARED BY: DR. JUNID
2. Discuss IaaS, PaaS, and SaaS cloud service models at different service levels.
Cloud computing delivers computing services over the internet on a pay-as-you-go basis. These
services are categorized into three major service models:
1. IaaS – Infrastructure as a Service
2. PaaS – Platform as a Service
3. SaaS – Software as a Service
These service models operate at different levels of abstraction and responsibility and are built
on top of each other as shown below:
SaaS ← Application Level (End-user)
PaaS ← Development Platform (Developer)
IaaS ← Infrastructure (Admin/System level)

3
PREPARED BY: DR. JUNID
Infrastructure as a Service (IaaS)
• Definition: Provides virtualized hardware resources like virtual machines, storage, and
networks over the internet.
• Control: User controls OS, storage, deployed apps; not the underlying infrastructure.
• Examples:
o Amazon EC2 (VMs)
o Amazon S3 (Storage)
o Google Compute Engine
o Microsoft Azure Virtual Machines
• Features:
o Virtual machines, storage, and networking
o Auto-scaling & load balancing
o Flexible & cost-efficient for developers and sysadmins

Platform as a Service (PaaS)

• Definition: Provides a complete development and deployment environment in the
cloud with tools and libraries.
• Control: User controls the deployed applications only; provider manages the platform
(OS, middleware, runtime).

4
PREPARED BY: DR. JUNID
 • Examples:
o Google App Engine
o Microsoft Azure App Services
o Heroku
o Red Hat OpenShift
• Features:
o Programming tools and APIs (Java, Python, .NET)
o Integrated development environments
o Automatic scaling, version control, testing

Software as a Service (SaaS)

• Definition: Provides ready-to-use software applications over the internet to end users.
• Control: User only uses the software; no control over platform or infrastructure.
• Examples:
o Gmail, Google Docs
o Microsoft Office 365
o Salesforce CRM
o Dropbox
• Features:
o Accessible via web browser
o No installation or maintenance needed
o Subscription-based or freemium model

3. Explain the design of data centers and their interconnection networks. [or] Describe the
modular data center concept, cooling system, inter-module networks, and data center
management issues.

A data center is a facility that houses computing resources such as servers, storage, and
networking equipment. It is essential for cloud services and enterprise-level computing. Data
centers are connected through high-speed interconnection networks for fast and reliable
communication.

Warehouse-Scale Data-Center Design

• Large data centers house 400,000 to 1 million servers.
• Built using economies of scale to reduce unit costs.
• Microsoft has over 100 data centers worldwide.
• A small data center may have 1,000 servers, but cost per unit is higher.

5
PREPARED BY: DR. JUNID
 • Example: Network cost in smaller DC is 7x more, storage cost is 5.7x more.

Construction Requirements
• Off-the-shelf servers with multi-core CPUs, DRAM, and disk drives.
• Example setup: 2,000 servers, each with 8 GB DRAM and 4×1 TB disks.
• First-level: Rack switches, Cluster-level: High-speed switches.
• Failures (hardware/software) are common; redundancy and replication are essential.
• Total disk storage is much higher than RAM but has lower bandwidth and higher latency.

Cooling System

• Data center rooms have raised floors to route cables and cooling air.
• Cooling is managed using CRAC (Computer Room Air Conditioning) units.
• Arrangement: Cold Aisles (cool air inlet) and Hot Aisles (exhaust).
• Cooling towers are used to pre-cool and reduce power usage.
• Efficient cooling reduces energy costs by up to 80% in newer setups.

Data-Center Interconnection Networks

Interconnection networks ensure efficient communication among thousands of servers.
Design Requirements:
• Low latency
• High bandwidth
• Low cost
• MPI (Message Passing Interface) support
• Fault tolerance

6
PREPARED BY: DR. JUNID
Application Traffic Support
• Network should support point-to-point and collective MPI communication.
• Needed for tasks like distributed file access, MapReduce, etc.
• High bisection bandwidth is required for parallel operations.

Network Expandability
• Must allow future growth without bottlenecks.
• Use of Fat-tree or Crossbar topologies with Ethernet switches.
• Modular design using data center containers supports scalability.
• Containers can be connected via external cabling and switches.

Fault Tolerance
• Provide redundancy in paths (multiple links) between servers.
• Avoid single point of failure.
• Must support graceful degradation and hot-swappable components.

Switch-Centric vs Server-Centric Designs

• Switch-centric: Switches manage all routing; servers remain unchanged.
• Server-centric: Servers are modified to assist in traffic routing.

Modular Data Centers in Shipping Containers

• Data centers now use truck-towable containers with blade servers.
• Example: SGI ICE Cube supports 46,080 cores or 30 PB per container.
• Easy to transport, scalable, and optimized for cooling and energy efficiency.
• Containers offer mobility, faster deployment, and reduced maintenance cost.

Container Construction
• Built with network, storage, and cooling inside the container.
• Steps: Build server ➝ rack ➝ full container ➝ testing.
• Requires weatherproofing and efficient cooling & airflow control.
• Useful in sectors like healthcare, defense, and remote cloud computing.

Interconnection of Modular Data Centers

• Containers are connected using inter-module networks like MDCube.
• BCube network inside each container handles internal communication.
• MDCube connects multiple containers using high-speed switches.

7
PREPARED BY: DR. JUNID
 • Forms a virtual hypercube at the container level for enhanced scalability.

Data Center Management Issues

A successful data center must address these management challenges:
Management Goal Description

User Satisfaction Provide consistent service for at least 30 years.

Controlled Info Flow Maintain high availability and streamlined traffic.

Multi-user Manageability Handle tasks like DB updates, maintenance, etc.

Scalability Easily grow with demand—storage, power, I/O, cooling.

Reliability Use failover, VM live migration, redundancy.

Low Cost Reduce costs for both users and providers.

Security Protect data from attacks, ensure data integrity.

Green IT Improve energy efficiency and reduce carbon footprint.

Marketplaces and Cost Management

• Factory-built containers reduce site-level installation work.
• Power densities exceed 1250 W/sq ft.
• Modular containers are suitable for rooftops, parking lots, etc.
• Must maintain redundancy, ensure future upgrades, and handle scalable growth.

8
PREPARED BY: DR. JUNID
4. Explain the Generic Cloud Architecture Design.

Design Goals of Cloud Platform

1. Scalability – Easily add or remove resources.
2. Virtualization – Allows sharing and better use of hardware.
3. Efficiency – Uses minimal resources to give maximum output.
4. Reliability – Cloud keeps working even if parts fail.
Enabling Technologies
• High-speed internet (broadband)
• Cheaper storage and powerful CPUs
• Virtualization (VMs)
• Web 2.0, SOA (Service Oriented Architecture)
• Auto billing, licensing tools
Generic Cloud Architecture
• Consists of many servers (physical/virtual)
• User requests go through a provisioning tool

9
PREPARED BY: DR. JUNID
 • Cloud includes:
o Storage (SANs, databases)
o Software (security, firewalls)
o Monitoring Tools (track usage, performance)
• Operated by third-party providers (Ex: AWS, Azure)
• Data centers located where electricity/cooling is cheap

5. Explain the Layered Cloud Architectural Development and Market-Oriented Cloud

Architecture.
Layered Cloud Architectural Development
Cloud architecture is developed in three layers:
1. Infrastructure Layer (IaaS):
• Virtualized compute, storage, and network resources.
• Handles provisioning and resource abstraction.
2. Platform Layer (PaaS):
• Provides software tools and environment for application development.
10
PREPARED BY: DR. JUNID
 • Ensures scalability, security, and dependability.
3. Application Layer (SaaS):
• Includes software modules for:
o Document editing
o Calendar/authentication
o CRM
o Financial transactions, etc.
Layer Dependency:
• Layers are built bottom-up (IaaS → PaaS → SaaS).
• Some services may span multiple layers.
Provider Workload:
• SaaS: Most workload for providers.
• PaaS: Medium.
• IaaS: Least workload.
Example: Amazon EC2 (IaaS), Salesforce CRM (SaaS + PaaS + IaaS).

Market-Oriented Cloud Architecture

To meet consumer demands and QoS (Quality of Service):

11
PREPARED BY: DR. JUNID
 • SLAs (Service-Level Agreements) define QoS.
• Market-based resource management balances supply and demand.
Main Components:
1. Users/Brokers: Submit service requests.
2. SLA Resource Allocator: Interfaces with users and assigns resources.
3. Service Request Examiner: Checks QoS and accepts/rejects requests.
4. Pricing Mechanism: Charges based on usage, peak/off-peak hours.
5. Accounting: Tracks actual resource usage.
6. VM Monitor: Monitors VM availability.
7. Dispatcher: Executes requests.
8. Service Request Monitor: Tracks execution.
VM Flexibility:
• Multiple VMs can run different OSes on one physical machine.
• VMs are isolated and dynamically scalable.

6. Explain the Virtualization Support and Disaster Recovery in Cloud Computing.

A key feature of cloud is virtualization, which hides the physical details from users.
Hardware Virtualization
• System virtualization software simulates hardware.
• Supports legacy systems, unusual apps, and multiple OSes.
• Development and deployment environments are the same, avoiding runtime errors.
12
PREPARED BY: DR. JUNID
 • Common tools: VMware, Microsoft, XenEnterprise.
• VMs offer:
o Full user control
o Custom OS
o Separation of user environments
• A virtual resource pool is created by virtualization managers.

Virtualization Support in Public Clouds

• AWS: Full VM flexibility.
• GAE (Google App Engine): Limited app-level virtualization.
• Microsoft Azure: Programming-level virtualization (.NET).
• Benefits:
o High availability
o Disaster recovery
o Load leveling
o Rich provisioning

Storage Virtualization for Green Data Centers

• Data centers consume ~3% of U.S. power.
• Green computing is needed.
• Virtualization + server consolidation saves power.
• Fortune 500 companies now adopt energy-efficient policies.

Virtualization for IaaS

Benefits of using VMs in IaaS:
1. Server consolidation.
2. Run legacy apps without interference.
3. Isolated environments increase security.
4. Performance isolation improves QoS guarantees.

VM Cloning for Disaster Recovery

Two methods:
1. Recover physical machine by another.
2. Recover VM using VM cloning.
• Cloning creates backup VMs on remote servers.
• Only one is active; others stay suspended.
• On failure, a snapshot is migrated to the clone.

13
PREPARED BY: DR. JUNID
 • Reduces downtime by 60%.
• Live migration is secured, only updating changes.

7. Explain the architectural design challenges in cloud computing.

Cloud architecture development faces several open challenges. These are identified as obstacles
and also opportunities for improvement. Armbrust et al. have highlighted some key
architectural challenges that affect cloud systems.
Challenge 1: Service Availability and Data Lock-in Problem
14
PREPARED BY: DR. JUNID
 • When a cloud is managed by a single provider, it becomes a single point of failure.
• Even with multiple data centers, the infrastructure might still share common software or
accounting systems.
Solution:
• Use multiple cloud providers for better High Availability (HA).
• To avoid DDoS attacks, some utility computing services offer quick scale-ups.
• APIs are still proprietary, leading to data lock-in.
Possible Fixes:
• Standardize APIs so SaaS developers can deploy across multiple clouds.
• Enables “surge computing”, where public clouds handle overflow from private clouds.

Challenge 2: Data Privacy and Security Concerns

• Clouds are mostly public networks, which increases the risk of attacks.
• Legal rules require providers to store data within national boundaries.
Traditional Attacks:
• Buffer overflows, DoS, spyware, malware, rootkits, Trojan horses, worms.
Cloud-Specific Threats:
• Hypervisor malware
• Guest hopping & hijacking
• VM rootkits
• Man-in-the-middle attacks during VM migration.
Solution:
• Use encryption, virtual LANs, firewalls, and packet filters.
• Apply security standards for both public and private cloud environments.

Challenge 3: Unpredictable Performance and Bottlenecks

• VMs can share CPUs/memory well, but I/O performance suffers.
• Example:
o 75 EC2 instances with STREAM → 1355 MB/s (good)
o Same 75 EC2 writing to local disk → 55 MB/s (bad)
Problem:
• I/O interference between VMs
• Data-intensive apps increase network traffic and data placement complexity
Solutions:
• Improve I/O virtualization, optimize interrupt handling
• Remove bottlenecks, upgrade weak links, discard weak servers
• Amazon CloudFront is an example of handling this with intelligent data distribution

15
PREPARED BY: DR. JUNID
Challenge 4: Distributed Storage and Widespread Software Bugs
• Databases in cloud apps are growing continuously
• Need for scalable, efficient SANs with high availability
• Data consistency in SAN-connected systems is a major issue
Bug Challenges:
• Bugs occur in production data centers and cannot be reproduced easily
Solutions:
• Use VMs to capture valuable info during runtime
• Simulators can also help with debugging at large scale if well-designed

Challenge 5: Cloud Scalability, Interoperability, and Standardization

Pricing Model Differences:
• Storage/network = Pay-per-byte
• Computation = Depends on VM or cycles
Examples:
• GAE: Charged per cycle, auto-scales up/down
• AWS EC2: Charged hourly per VM, even if idle
Standardization via OVF (Open Virtualization Format):
• Portable, secure, cross-platform VM format
• Doesn’t depend on specific OS or virtualization tool
• Can include multiple VMs per appliance
Needed Improvements:
• Hypervisor-agnostic VMs
• Cross-platform live migration
• Support for legacy hardware
• Standardized VM packaging and transport
• Load balancing across x86 Intel and AMD

Challenge 6: Software Licensing and Reputation Sharing

Licensing Issues:
• Traditional software licensing doesn’t fit cloud models
• Commercial licensing is not ideal for utility computing
Opportunities:
• Open-source software may remain dominant
• Pay-per-use or bulk-use licensing models can be adopted
Reputation Issues:

16
PREPARED BY: DR. JUNID
 • Bad behavior by one customer (e.g., spam) may affect the reputation of the entire cloud
• EC2 IP addresses can be blacklisted, affecting others
Solutions:
• Create reputation-guarding services (like trusted email services)
• Legal liability transfer must be addressed via SLA agreements

8. Describe Google App Engine (GAE). Explain its architecture, functional modules, and
applications in detail.
Google App Engine (GAE) is a Platform as a Service (PaaS) offering by Google. It enables
developers to build, deploy, and scale web applications without managing the underlying
infrastructure.
Google has the world’s largest search engine and operates hundreds of data centers globally
with more than 460,000 servers. These are used for massive-scale cloud applications with high
availability (HA). GAE is built upon this powerful infrastructure and supports various cloud and
web applications.

Google Cloud Infrastructure

Google pioneered cloud development by leveraging its massive data centers, which power
services like:
• Gmail
• Google Docs
• Google Earth
It achieved breakthroughs with technologies such as:
• Google File System (GFS) – for storing large-scale data
• MapReduce – for processing large data sets
• BigTable – for structured data storage
• Chubby – for distributed locking services
GAE was launched in 2008 to support scalable (elastic) web apps. It allows third-party
developers to run their apps directly on Google’s cloud infrastructure across many data centers.

GAE Architecture
GAE architecture consists of the following core components:
1. GFS – For large-scale data storage.
2. MapReduce – For large-scale data processing.
3. BigTable – A distributed structured storage system.
4. Chubby – A distributed locking service.

17
PREPARED BY: DR. JUNID
 • These technologies power cloud services internally. Applications run on Google-
managed data centers with thousands of servers organized into clusters.
• Users interact through web interfaces, and third-party developers build apps using GAE
which runs in the same environment as Google’s own applications.
• GAE abstracts server management, enabling developers to focus only on coding, not
infrastructure.
• GAE provides a frontend application framework similar to ASP, J2EE, JSP, and
supports Python and Java programming environments.

Functional Modules of GAE

GAE consists of the following five functional modules:
1. Datastore:
o Object-oriented, distributed, and structured data storage.
o Built using BigTable technology.
o Provides secure data management and query support.
2. Application Runtime Environment:
o Scalable execution environment for web apps.
o Supports Python and Java for application development.
3. Software Development Kit (SDK):
o Used for local development and testing.
o Allows developers to run and test applications locally before uploading to GAE.
4. Administration Console:
o Provides tools to manage application lifecycle and deployment.
o Does not deal with physical resource management.

18
PREPARED BY: DR. JUNID
 5. GAE Web Service Infrastructure:
o Special APIs for storage and network resource management.
o Offers features like user authentication (via Gmail accounts) and email services.
GAE is free for Gmail users within a quota. If usage exceeds limits, users are instructed on how
to pay for the services.

GAE: Language and Platform Support

• GAE supports applications written in Python, Java, and Ruby.
• It is a PaaS platform only, unlike Amazon which provides both IaaS + PaaS.
• Developers use supported tools/languages but do not control or manage the underlying
infrastructure.
• This is similar to Microsoft Azure, which supports .NET applications.

GAE Applications
Google App Engine supports several well-known applications:
Examples:
• Google Search
• Gmail
• Google Docs
• Google Earth
All of these apps run on Google’s internal cloud infrastructure managed within large-scale data
centers.
GAE supports third-party developers to:
• Build and run web apps on the Google infrastructure.
• Store application-specific data.
• Use backend persistent storage with support for queries, sorting, and
transactions similar to traditional DBMS.
Bonus GAE Feature:
• Web apps can use Gmail accounts for authentication and email services.
• This eliminates the need for custom user management modules.

9. Describe Amazon Web Services (AWS). Explain its architecture, major services, and
features.
Amazon Web Services (AWS) is a leading public cloud platform that follows the Infrastructure
as a Service (IaaS) model. It allows developers and companies to access computing
infrastructure like servers, storage, and networking via the internet.

19
PREPARED BY: DR. JUNID
AWS provides flexible, scalable, and cost-efficient resources, which can be shared safely using
Virtual Machines (VMs).

AWS Architecture
The architecture of AWS is built using several components. Some key components are:
Component Function

EC2 (Elastic Compute Cloud) Provides virtual machines (VMs) to run cloud applications.

S3 (Simple Storage Service) Offers object-oriented storage for saving data in the cloud.

EBS (Elastic Block Store) Provides block-level storage for traditional applications.

SQS (Simple Queue Service) Reliable message queue system used for communication between process

SNS (Simple Notification Service) Sends messages or alerts to distributed systems or users.
These services allow users to build, deploy, and scale cloud-based applications.

Working of AWS Components

• EC2 is used to launch and manage virtual servers.
• S3 allows users to store any type of data as objects (files, images, etc.).
• EBS acts like a hard drive for EC2 instances and stores structured data.
• SQS ensures that messages are stored safely even if the receiver is offline.
• Users can access AWS services via SOAP-enabled web browsers or client programs.

20
PREPARED BY: DR. JUNID
Extra AWS Services
AWS also provides additional services in 12 application areas. Some of them include:
Service Purpose

CloudWatch Monitoring tool for AWS resources (e.g., CPU, disk, network).

ELB (Elastic Load Balancing) Distributes incoming traffic to healthy EC2 instances.

Auto Scaling Automatically adjusts the number of EC2 instances based on load.

RDS (Relational Database Service) Provides managed SQL databases in the cloud.

Elastic MapReduce Big data processing framework similar to Hadoop.

Import/Export Allows transfer of large data by shipping physical disks.

CloudFront Content Delivery Network (CDN) for faster content access.

DevPay Manages billing and online payments for apps built on AWS.

FPS (Flexible Payment Service) Lets developers charge users using Amazon’s payment system.

FWS (Fulfillment Web Service) Allows merchants to use Amazon’s logistics to fulfill orders.

Advantages of AWS Platform

Flexibility: Developers can use AWS to run various types of applications without owning physical
hardware.
Support for Startups & SMEs: Small and medium businesses can move their entire
infrastructure to AWS to support users globally and earn profits.
Monitoring & Reliability:
• CloudWatch monitors EC2 and other resources.
• ELB and Auto Scaling help handle high traffic and maintain high availability.
Billing & Licensing:
• AWS offers services like DevPay and FPS for easy billing and online transactions.
• Customers can use their existing Amazon login and payment info for purchases.
Cluster Computing Support:
• In 2010, AWS introduced MPI clusters and Cluster Compute Instances.

21
PREPARED BY: DR. JUNID
 • These use hardware-assisted virtualization, improving performance over para-
virtualized instances.

10. Discuss a set of cloud services provided by Microsoft Azure.

Microsoft Azure is a cloud computing platform launched in 2008, built over Microsoft’s global
data centers. It provides Infrastructure as a Service (IaaS), Platform as a Service (PaaS),
and Software as a Service (SaaS) to develop, manage, and host applications.

Cloud Services in Microsoft Azure

1. Live Services
• Enables users to access Microsoft Live applications.
• Supports real-time data sharing and synchronization across multiple machines.
2. .NET Services
• Supports application development using Microsoft’s .NET framework.
• Applications can be developed locally and then deployed on Azure cloud servers.
3. SQL Azure
• A cloud-based relational database service.
• Provides easy access to SQL Server functionalities over the cloud, with scalability and
high availability.
4. SharePoint Services
• Offers a scalable platform for building custom business applications.
• Supports web services with tools for collaboration, data sharing, and content
management.
5. Dynamic CRM Services
• Provides a customer relationship management (CRM) platform.
• Helps in finance, marketing, sales, and promotional activities using cloud-based tools.

22
PREPARED BY: DR. JUNID
Integration and Support
• Azure services can integrate with traditional Microsoft products like:
o Windows Live, Office Live
o Exchange Online, SharePoint Online
o Dynamic CRM Online
• Uses standard protocols like SOAP and REST for cloud communication.
• Supports integration with third-party platforms and clouds.

Development Tools
• Azure SDK (Software Development Kit) allows:
o Local development and testing
o Easy deployment and debugging of cloud apps on Windows machines

11. Write short notes on global exchange of cloud resources.

To serve a global user base, cloud providers like Amazon, Google, and Microsoft set up data
centers across the world. However, hosting and scaling services across these geographically
distributed centers faces challenges in meeting QoS (Quality of Service) needs of users.

1. Multiple Data Centers

23
PREPARED BY: DR. JUNID
 • Cloud providers (IaaS) have data centers in different countries for reliability and
redundancy.
• Example: Amazon has centers in US East, US West, and Europe.
2. Challenges Faced
• Customers must choose a hosting region manually, without knowing where most users
are.
• This leads to poor performance for distant users and difficulty maintaining QoS globally.
3. Need for Federation
• No single provider can cover the entire globe.
• SaaS providers want to use multiple cloud providers for better regional performance.

InterCloud Architecture
A proposed solution that:
• Federates multiple cloud providers.
• Enables dynamic resource sharing and scaling.
• Supports brokering and SLA-based negotiation between providers and customers.

Cloud Exchange (CEx)

• Acts as a marketplace between service providers and consumers.
• Supports auction-based and market-driven pricing.
• Handles secure financial transactions via a virtual banking system.
• Helps match resource demand and supply across clouds.

Benefits
• Global scalability
• Cost-effective service delivery
• High QoS and reliability
• On-demand resource expansion
• Efficient workload migration

24
PREPARED BY: DR. JUNID
 25
PREPARED BY: DR. JUNID