0% found this document useful (0 votes)

17 views284 pages

Cloud - Completeslides 2

This presentation covers cloud computing concepts, including architecture, deployment models, and services. It provides insights into scalability, cost-efficiency, and industry use cases of cloud technology. Beneficial for computer science students, IT professionals, and researchers exploring cloud soluti

Uploaded by

jsingh5be22

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

17 views284 pages

Cloud - Completeslides 2

Uploaded by

jsingh5be22

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 284

Course on Cloud Computing

Introduction

Dr. Gitanjali Chandwani Manocha

Assistant Professor
Department of Electronics & Communication Engg
Thapar Institute of Engineering and Technology, Patiala
Motivation
Motivation

How it Cloud Computing has grown exponentially in recent years

Evolved

Introduction

Basics of
Cloud
Computing

Introduction Hype Cycle is a branded graphical depiction developed and used

Basics of by Gartner to represent common pattern that arises with each
Cloud
Computing
new technology or other innovation
Further
Reading
Hype Cycle for Clould
Motivation

How it
Computing 2011
Evolved

Introduction

Basics of
Cloud
Computing

Further
Reading
Hype Cycle for Clould
Motivation

How it
Computing 2018
Evolved

Introduction

Basics of
Cloud
Computing

Further
Reading
Overview of Cloud Computing
Motivation

How it
Evolved

Introduction
Overview of Cloud
Computing
1 Resources: Servers, network, service and application
Resources
What is Cloud
2 Essential characteristics: On-demand self-service,
Computing
Why Cloud
Computing
broad network access, resource pooling, rapid
Basics of
elasticity, measure service
Cloud
Computing 3 Deployment models: public, private, community and
Further hybrid cloud
Reading
4 Delivery models: XaaS (X as a service); X: Software,
Platform, Infrastructure
5 Infrastructure: distributed infrastructure, resource
virtualization, autonomous system
Resources
Motivation

How it
Evolved

Introduction
Server: compute and storage
Overview of Cloud
Computing
Data centers
Resources
What is Cloud
Computing
Network
Why Cloud
Computing Service
Basics of
Cloud Application
Computing

Further
Reading

Facebook data center

Datacenters
Motivation

How it
Evolved

Introduction
Overview of Cloud
Computing
Resources
What is Cloud
Computing
Why Cloud
Computing

Basics of
Cloud
Computing

Further
Reading

Facebook data center-Inside View

Resources
Motivation

How it
Evolved

Introduction
Overview of Cloud
Computing
Resources
What is Cloud
Computing
Why Cloud Server: compute and storage
Computing

Basics of
Data centers
Cloud
Computing Network
Further
Reading
Service
Application
What is Cloud Computing
Motivation

How it
Evolved On-demand delivery of computing service over
Introduction
Internet:
Overview of Cloud
Computing Servers
Resources
What is Cloud
Storage
Computing
Why Cloud
Processing Power
Computing
Network
Basics of
Cloud IT infrastructure
Computing Development platforms
Further Database
Reading
The term cloud has historically been used in the
telecommunications industry as an abstraction of the
network in system diagrams.
What is Cloud Computing
Motivation

How it
Evolved

Introduction National Institute of Standards and Technology (NIST)

Overview of Cloud
Computing
Resources
Cloud computing is a model for enabling ubiquitous,
What is Cloud
Computing
convenient, on-demand network access to a shared pool
Why Cloud
Computing of configurable computing resources (e.g., networks,
Basics of servers, storage, applications, and services) that can be
Cloud
Computing rapidly provisioned and released with minimal
Further
Reading
management effort or service provider interaction.

Gartners:
Cloud computing is a style of computing in which scalable
and elastic IT-enabled capabilities are delivered as a
service using internet technologies.
What is Cloud Computing
Motivation

How it
Evolved
Buyya et al.
Introduction
Overview of Cloud
Computing
A cloud is a type of parallel and distributed system
Resources
What is Cloud
consisting of a collection of interconnected and
Computing
Why Cloud virtualized computers that are dynamically provisioned
Computing

Basics of
and presented as one or more unified computing resources
Cloud
Computing
based on service-level agreements established through
Further
negotiation between the service provider and consumers
Reading

Oracle
cloud computing is renting instead of buying your IT.
Rather than investing heavily in databases, software, and
equipment, companies are opting to access their compute
power via the internet and pay for it as they use it
Why Cloud Computing
Motivation On Premises Cloud
How it
Evolved Buy/own servers or data Huge saving in CAPEX-no
Introduction centers need to set IT infrastructure
Overview of Cloud
Computing
Resources Space to set up IT Offsite data center
What is Cloud
Computing infrastructure
Why Cloud
Computing
User requires to maintain Cloud provider maintains
Basics of
Cloud hardware and software; data centers; provides
Computing
manpower to maintain it self-service interface to
Further
Reading users
upfront payment Pay-per-use go
Expanding infrastructure: Illusion of infinite resources
costly & time consuming
Does not provide elasticity Provides elasticity
Planning time high On Demand-witin few mins
Basics of Cloud Computing
Motivation

How it
Evolved

Introduction Cloud services are (almost alway) delivered remotely

Basics of from off-site data centers
Cloud
Computing Cloud consists of physical machines in the data centers
Basics of Cloud
Computing of cloud providers.
Essential
Characteristics Cloud services are delivered by virtualization-creating
Use Cases
virtual machines
Further
Reading From user perspective virtualized hardware is identical
to local physical hardware
Software that performs virtulization is called hypervisor.
Cloud economics is favorable in general
Flexible charging plan-charged for service used
Organization of every type, size can have access to
similar IT infrastructure
Basics of Cloud Computing
Motivation

How it
Evolved

Introduction

Basics of
Cloud
Computing
Users can manage cloud services through a console
Basics of Cloud
Computing Provide mobility
Essential
Characteristics
Use Cases
24×7 availability
Further
Reading
Tolerant to failures
Cloud computing services offered by 3rd party
Data is under control of cloud provider
security of data
Essential Characteristics
Motivation
1 On-demand self-service.
How it
Evolved User can select the resources needed
Introduction User can access anytime anywhere
Basics of Users can add/remove services when ever they want
Cloud
Computing The service is accessible via a Web browser or a Web
Basics of Cloud
Computing services API
Essential
Characteristics
Use Cases
2 Broad network access: The nature of the cloud computing
Further
should support all the standard protocols & devices.
Reading
Any network & any web-enabled device
3 Resource pooling: resources are shared among multiple
users and are allocated dynamically in a multi-tenant
environment
4 Rapid elasticity- ability to scale up (or scale down) resources
whenever required
5 Measured services: Resource usage can be monitored,
controlled, and reported, providing transparency for both the
provider and consumer of the utilized service.
Use Cases
Motivation

How it
Evolved

Introduction
Business-agile, reduce cost, instantly scale and deployed
Basics of
Cloud globally in minutes
Computing
Basics of Cloud
Computing
1 Healthcare companies: to develop personalized
Essential
Characteristics treatments for patients
Use Cases

Hospitality industry, Production industry, Financial

industry and many many more
Further Reading
Motivation

How it
Evolved

Introduction

Basics of
Cloud
Computing 1 Kai Hwang , Jack Dongarra , Geoffrey C. Fox Distributed and Cloud Computing: From Parallel Processing
to the Internet of Things. Morgan Kauffman 2011
Further
Reading 2 Rajkumar Buyya, James Broberg and Goscinski Author Name, Cloud Computing Principles and
Paradigms, John Wiley and Sons 2012, Second Edition
3 Gerard Blokdijk, Ivanka Menken,The Complete Cornerstone Guide to Cloud Computing Best Practices,
Emereo Pvt Ltd, 2009, Second Edition
4 Anthony Velte, Toby Velte and Robert Elsenpeter , Cloud Computing: A practical Approach Tata
McGrawHill, 2010, Second Edition
5 Judith Hurwitz, Robin Bllor, Marcia Kaufmann, Fern Halper, Cloud Computing for Dummies, 2009, Third
Edition
Motivation

How it
Evolved

Introduction

Basics of
Cloud
Computing

Further
Reading
Thank You
Course on Cloud Computing

Evolution of Cloud Computing

Dr. Gitanjali Chandwani Manocha

Assistant Professor
Department of Electronics & Communication Engg
Thapar Institute of Engineering and Technology, Patiala
Trends in Computing
Trends in
Computing
Distributed
computing
Cluster computing
Grid Computing
Utility computing
Cloud Computing
Further Reading Centralized Comuting
Parallel Computing
Distributed computing
Cluster Computing
Grid Computing
Utility Computing
Cloud Computing
Centralized Computing
Trends in
Computing
Distributed
computing
Cluster computing
Grid Computing Cloud computing overlaps with distributed, centralized, and
Utility computing
Cloud Computing parallel computing.
Further Reading

Centralized computing
Computer resources are centralized in one physical
system
Centralized Computing
Trends in
Computing
Distributed
computing
Cluster computing
Grid Computing Cloud computing overlaps with distributed, centralized, and
Utility computing
Cloud Computing parallel computing.
Further Reading

Centralized computing
Computer resources are centralized in one physical
system
Resources (processors, memory, and storage) are fully
shared and tightly coupled within one integrated OS.
Centralized Computing
Trends in
Computing
Distributed
computing
Cluster computing
Grid Computing Cloud computing overlaps with distributed, centralized, and
Utility computing
Cloud Computing parallel computing.
Further Reading

Centralized computing
Computer resources are centralized in one physical
system
Resources (processors, memory, and storage) are fully
shared and tightly coupled within one integrated OS.
Many data centers and mainframes are centralized
systems, but they are used in parallel, distributed, and
cloud computing applications
Parallel computing
Trends in
Computing
Distributed
computing
Cluster computing
Grid Computing
Utility computing
Cloud Computing The field of parallel computing overlaps with distributed
Further Reading
computing to a great extent

Parallel computing
Processors are either tightly coupled with centralized
shared memory or loosely coupled with distributed
memory.
Interprocessor communication is accomplished through
shared memory or via message passing.
Distributed computing
Trends in
Computing
Distributed
computing
Cluster computing
Clouds are essentially large distributed computing facilities
Grid Computing
Utility computing
that make available their services to third parties on
Cloud Computing
Further Reading
demand.
Distributed computing
It is a field of computer science that studies distributed
systems
Distributed system: components are located on
different networked computers
Information exchange in such a distributed system is
accomplished through message passing

Parallel computing can be considered as a subset of

distributed computing
Cluster computing
Trends in
Computing
Distributed
computing
Cluster computing
What is it?
Grid Computing
Utility computing Many of the computers connected (closely) on a network
Cloud Computing
Further Reading (LAN) and they perform like a single entity.

Low-cost alternative to the use of mainframes and

supercomputers
Cluster computing
Trends in
Computing
Distributed
computing
Cluster computing
What is it?
Grid Computing
Utility computing Many of the computers connected (closely) on a network
Cloud Computing
Further Reading (LAN) and they perform like a single entity.

Low-cost alternative to the use of mainframes and

supercomputers
Commodity machines could connected by a
high-bandwidth network and controlled by specific
software tools that manage them as a single system.
Cluster computing
Trends in
Computing
Distributed
computing
Cluster computing
What is it?
Grid Computing
Utility computing Many of the computers connected (closely) on a network
Cloud Computing
Further Reading (LAN) and they perform like a single entity.

Low-cost alternative to the use of mainframes and

supercomputers
Commodity machines could connected by a
high-bandwidth network and controlled by specific
software tools that manage them as a single system.
Built by machines, they were cheaper than mainframes
and made high-performance computing available to a
large number of groups,
Cluster computing
Trends in
Computing
Distributed
computing
Cluster computing
What is it?
Grid Computing
Utility computing Many of the computers connected (closely) on a network
Cloud Computing
Further Reading (LAN) and they perform like a single entity.

Low-cost alternative to the use of mainframes and

supercomputers
Commodity machines could connected by a
high-bandwidth network and controlled by specific
software tools that manage them as a single system.
Built by machines, they were cheaper than mainframes
and made high-performance computing available to a
large number of groups,
Clusters are easy to extend if more power is required
Grid Computing
Trends in
Computing
Distributed
Evolution of Grid
computing
Cluster computing Clusters became quite common resources
Grid Computing
Utility computing
Cloud Computing
Further Reading

Grid computing is a combination of resources from multiple

administrative domains to reach a common target
Grid Computing
Trends in
Computing
Distributed
Evolution of Grid
computing
Cluster computing Clusters became quite common resources
Grid Computing
Utility computing
Cloud Computing
They were often underutilized
Further Reading
Demand for computational power beyond the capability of
single clusters

Grid computing is a combination of resources from multiple

administrative domains to reach a common target
Computers do not have to be in the same physical
location and operated independently.
Grid Computing
Trends in
Computing
Distributed
Evolution of Grid
computing
Cluster computing Clusters became quite common resources
Grid Computing
Utility computing
Cloud Computing
They were often underutilized
Further Reading
Demand for computational power beyond the capability of
single clusters
Improvements in networking and the Internet made possible
long-distance, high-bandwidth connectivity.

Grid computing is a combination of resources from multiple

administrative domains to reach a common target
Computers do not have to be in the same physical
location and operated independently.
Each computer is a distinct computer-different OS and
hardware
Grid Computing
Trends in
Computing
Distributed
Evolution of Grid
computing
Cluster computing Clusters became quite common resources
Grid Computing
Utility computing
Cloud Computing
They were often underutilized
Further Reading
Demand for computational power beyond the capability of
single clusters
Improvements in networking and the Internet made possible
long-distance, high-bandwidth connectivity.

Grid computing is a combination of resources from multiple

administrative domains to reach a common target
Computers do not have to be in the same physical
location and operated independently.
Each computer is a distinct computer-different OS and
hardware
Different from "large cluster its scale was nationwide or
even worldwide
Utility computing
Trends in
Computing
Distributed
computing
What it is and it is not?
Cluster computing
Grid Computing it is not a new paradigm of computing infrastructure
Utility computing
Cloud Computing
Further Reading
it is a business model in which computing resources,
such as computation and storage, are packaged as
metered services similar to a physical public utility
Utility computing is typically implemented using other
computing infrastructure (e.g. Grids) with additional
accounting and monitoring services.

Users assign a utility value to their job

Utility computing
Trends in
Computing
Distributed
computing
What it is and it is not?
Cluster computing
Grid Computing it is not a new paradigm of computing infrastructure
Utility computing
Cloud Computing
Further Reading
it is a business model in which computing resources,
such as computation and storage, are packaged as
metered services similar to a physical public utility
Utility computing is typically implemented using other
computing infrastructure (e.g. Grids) with additional
accounting and monitoring services.

Users assign a utility value to their job

shared systems are viewed as a marketplace, where
users compete for resources based on the perceived
utility or value of their jobs.
Utility computing
Trends in
Computing
Distributed
computing
What it is and it is not?
Cluster computing
Grid Computing it is not a new paradigm of computing infrastructure
Utility computing
Cloud Computing
Further Reading
it is a business model in which computing resources,
such as computation and storage, are packaged as
metered services similar to a physical public utility
Utility computing is typically implemented using other
computing infrastructure (e.g. Grids) with additional
accounting and monitoring services.

Users assign a utility value to their job

shared systems are viewed as a marketplace, where
users compete for resources based on the perceived
utility or value of their jobs.
shared systems are not geographically co-located.
Cloud Computing
Trends in
Computing
Distributed
computing
Cluster computing
Definition overlaps with technologies such as Grid, Utility
Grid Computing
Utility computing
and distributed computing in general
Cloud Computing
Further Reading
Cluster Computing is focused on traditional non-service
applications.
Cloud Computing
Trends in
Computing
Distributed
computing
Cluster computing
Definition overlaps with technologies such as Grid, Utility
Grid Computing
Utility computing
and distributed computing in general
Cloud Computing
Further Reading
Cluster Computing is focused on traditional non-service
applications.
It is indeed evolved out of Grid Computing
Cloud Computing
Trends in
Computing
Distributed
computing
Cluster computing
Definition overlaps with technologies such as Grid, Utility
Grid Computing
Utility computing
and distributed computing in general
Cloud Computing
Further Reading
Cluster Computing is focused on traditional non-service
applications.
It is indeed evolved out of Grid Computing
Grid computing enables aggregation of distributed
resources and transparently access to them
Cloud Computing
Trends in
Computing
Distributed
computing
Cluster computing
Definition overlaps with technologies such as Grid, Utility
Grid Computing
Utility computing
and distributed computing in general
Cloud Computing
Further Reading
Cluster Computing is focused on traditional non-service
applications.
It is indeed evolved out of Grid Computing
Grid computing enables aggregation of distributed
resources and transparently access to them
Due to shift in focus from an infrastructure that delivers
storage and compute resources ( case in Grids) to one
that is economy based aiming to deliver more abstract
resources and services (case in Clouds).
Cloud Computing
Trends in
Computing
Distributed
computing
Cluster computing
Definition overlaps with technologies such as Grid, Utility
Grid Computing
Utility computing
and distributed computing in general
Cloud Computing
Further Reading
Cluster Computing is focused on traditional non-service
applications.
It is indeed evolved out of Grid Computing
Grid computing enables aggregation of distributed
resources and transparently access to them
Due to shift in focus from an infrastructure that delivers
storage and compute resources ( case in Grids) to one
that is economy based aiming to deliver more abstract
resources and services (case in Clouds).
It is a broader concept than utility computing; Cloud
Computing lies at the large-scale side.
Futher Reading
Trends in
Computing
Distributed
computing
Cluster computing
Grid Computing
Utility computing
Cloud Computing
Further Reading

1 Foster, Ian, et al. "Cloud computing and grid computing

360-degree compared." 2008 grid computing environments
workshop. Ieee, 2008.
Trends in
Computing
Distributed
computing
Cluster computing
Grid Computing
Utility computing
Cloud Computing
Further Reading

Thank You
Course on Cloud Computing

Deployment and Delivery Models

Dr. Gitanjali Chandwani Manocha

Assistant Professor
Department of Electronics & Communication Engg
Thapar Institute of Engineering and Technology, Patiala
Overview of Cloud Computing
Overview of
Cloud
Computing

Deployment
Models
1 Resources: Servers, network, service & application X
Delivery
Models

Further
Reading
Overview of Cloud Computing
Overview of
Cloud
Computing

Deployment
Models
1 Resources: Servers, network, service & application X
Delivery
Models 2 Defining attribute: On-demand self-service, broad
Further
Reading network access, resource pooling, rapid elasticity,
measure service X
3 Deployment models: public, private, community and
hybrid cloud
4 Delivery models: XaaS (X as a service); X: Software,
Platform, Infrastructure
Overview of Cloud Computing
Overview of
Cloud
Computing

Deployment
Models
1 Resources: Servers, network, service & application X
Delivery
Models 2 Defining attribute: On-demand self-service, broad
Further
Reading network access, resource pooling, rapid elasticity,
measure service X
3 Deployment models: public, private, community and
hybrid cloud
4 Delivery models: XaaS (X as a service); X: Software,
Platform, Infrastructure
5 Infrastructure: distributed infrastructure, resource
virtualization, autonomous system
Public Cloud
Overview of
Cloud
Computing

Deployment The cloud infrastructure is provisioned for open use by

Models
Public Cloud the general public on pay-as-you-go manner
Private Cloud
Community Cloud
Infrastructure exists on the premises of the cloud
Delivery
Models
provider
Further Services are provided by a third-party cloud provider
Reading
Cloud resources are shared among different users

Suited for
Users who want cloud infrastructure for development
and testing of applications
Hosting applications in cloud to serve large workload
without upfront cost
Private Cloud
Overview of
Cloud
Computing The cloud infrastructure is provisioned for exclusive use
Deployment by a single organization
Models
Public Cloud It is client owned and managed, and its access is
Private Cloud
Community Cloud limited to the owning clients and their partners
Delivery
Models Its deployment is not meant to sell capacity over the
Further Internet through publicly accessible interfaces
Reading
It is built within the domain of an intranet owned by a
single organization

Suited for
Applications where security is very important
Institutions such as governments and banks that have
high security, privacy, and regulatory concerns prefer to
build and use their own private clouds.
Users who want to have tight control over sensitive data
Private Cloud
Overview of
Cloud
Computing

Deployment
Models
Public Cloud
Private Cloud
Virtual Private Cloud (VPC)- cloud infrastructure is
Community Cloud provisioned for exclusive use by a single organization but is
Delivery
Models
hosted by a third party
Further Cloud provider dedicates a set of computing resources
Reading
for a specific customer
Cloud provider maintains the infrastructure
Provides service over Virual Private Network (VPN)
Virtual refers to the fact that the physical infrastructure
constituting the private cloud is situated off premise.
Community Cloud
Overview of
Cloud
Computing

Deployment
Models
The cloud infrastructure is provisioned for exclusive use
Public Cloud
Private Cloud
by a specific community of consumers that have shared
Community Cloud
concerns (e.g., mission, security requirements, policy,
Delivery
Models and compliance considerations).
Further
Reading
It may be owned, managed, and operated by one or
more of the organizations in the community, a third
party, or some combination of them, and it may exist on
or off premises

Suited for
Organizations that want access same data/application
and want cloud cost to be shared with the larger group
Hybrid Cloud
Overview of
Cloud
Computing

Deployment
Models The cloud infrastructure is a composition of two or more
Public Cloud
Private Cloud distinct cloud infrastructures (private, community, or
Community Cloud
public)
Delivery
Models Private cloud is supplemented with computing facility
Further from public cloud
Reading
This approach of temporarily renting capacity to handle
spike in load is known as cloud bursting

Suited for
Users who want to take advantage of secured hosting
on a private cloud and cost saving due to shared
hosting in public cloud
Delivery models
Overview of Services can be classified into 3 major categories:
Cloud
Computing Software-as-a-Service (SaaS), Platform-as-a-Service
Deployment (PaaS),Infrastructure-as-a-Service (IaaS)
Models

Source [3]
Delivery models
Overview of Services can be classified into 3 major categories:
Cloud
Computing Software-as-a-Service (SaaS), Platform-as-a-Service
Deployment (PaaS),Infrastructure-as-a-Service (IaaS)
Models

Delivery
Service: A service is a mechanism that is capable of
Models providing one or more functionalities, which it is possible to
SaaS
PaaS use in compliance with provider-defined restrictions and
IaaS
rules and through an interface .
Further
Reading Platform: A platform is a fundamental computer system that
includes hardware equipment, operating systems, and, in
some cases, application development tools and user inter-
faces on which applications can be deployed and executed.
Infrastructure: Infrastructure refers to underlying physical
components that are required for a system to perform its
functionalities. In information systems, these components
can contain processors, storage, network equipment, and, in
some cases, database management systems and operating
systems.
Source [3]
Software as a Service
Overview of
Cloud
Computing

Deployment
Models

Delivery
Models
The capability provided to the consumer is to use the
SaaS
PaaS
provider’s applications running on a cloud infrastructure
IaaS
The applications platform independent and are
Further
Reading accessible from various client devices through
interface such as a web browser or a program interface.
Consumers are increasingly shifting from locally
installed computer programs to on-line software
services that offer the same functionally as it alleviates
the burden of software maintenance for customers
Software as a Service
Overview of
Cloud
Computing

Deployment
Models

Delivery The consumer does not manage or control the

Models
SaaS underlying cloud infrastructure:
PaaS
IaaS Network
Further Servers
Reading
Operating systems
Storage
Individual application capabilities

Limited control over user-specific application

configuration settings.
Platform as a Service
Overview of
Cloud
Computing

Deployment
Models It offers an environment on which developers develop,
Delivery
Models
deploy, configure and manage the application
SaaS
PaaS
applications
IaaS
Multiple programming models and specialized services
Further
Reading (e.g., data access, authentication, and payments) are
offered to build new applications.
Users need to not know how many processors or how
much memory that applications will be using.
Example Google AppEngine - offers a scalable
environment for developing and hosting Web
applications
Platform as a Service
Overview of
Cloud
Computing

Deployment
Models

Delivery
Models The consumer does not manage or control the
SaaS
PaaS
underlying cloud infrastructure:
IaaS
Network
Further
Reading
Servers
Operating systems
Storage
Control over the deployed applications and possibly
configuration settings for the application-hosting
environment.
Infrastructure as a Service
Overview of
Cloud
Computing

Deployment
Models
Offers virtualized resources on demand
Delivery
Models The capability provided to the user is to provision
SaaS
PaaS processing, storage, networks, and other fundamental
IaaS
computing resources
Further
Reading User is able to deploy and run operating systems and
applications
The consumer does not manage or control the underlying
cloud infrastructure
User has control over operating systems, storage, and
deployed applications and possibly limited control of select
networking components (e.g., host firewalls)
Delivery models
Overview of
Cloud
Computing

Deployment
Models

Delivery
Models
SaaS
PaaS
IaaS

Further
Reading

The Cloud Computing Reference Model. Courtesy [1]

Cloud Stack
Overview of
Cloud
Computing

Deployment
Models

Delivery
Models
SaaS
PaaS
IaaS

Further
Reading
Further Reading
Overview of
Cloud
Computing

Deployment
Models

Delivery
Models

Further
Reading
1 Buyya, Rajkumar, Christian Vecchiola, and S. Thamarai Selvi. Mastering cloud computing: foundations
and applications programming. Newnes, 2013.
2 Mell, Peter, and Tim Grance. "The NIST definition of cloud computing." (2011).
3 Jula, Amin, Elankovan Sundararajan, and Zalinda Othman. "Cloud computing service composition: A
systematic literature review." Expert systems with applications 41.8 (2014): 3809-3824.
Overview of
Cloud
Computing

Deployment
Models

Delivery
Models

Further

Thank you
Reading
Course on Cloud Computing

Cloud Infrastructure and Virtualization

Dr. Gitanjali Chandwani Manocha

Assistant Professor
Department of Electronics & Communication Engg
Thapar Institute of Engineering and Technology, Patiala
Overview of Cloud Computing
Introduction

Infrastructure

Layered
Architecture

Virtualization
1 Resources: Servers, network, service & application X
Further 2 Defining attribute: On-demand self-service, broad
Reading
network access, resource pooling, rapid elasticity,
measure service X
3 Deployment models: public, private, community and
hybrid cloud X
4 Delivery models: XaaS (X as a service); X: Software,
Platform, Infrastructure X
5 Infrastructure: distributed infrastructure, resource
virtualization, autonomous system
Infrastructure
Introduction

Infrastructure

Layered
Architecture

Virtualization
Distributed infrastructure: Servers/datacenters need
Further
Reading not be located in the same physical location as user.
Services can be aquired from Anywhere
Resource Virtualization: Virtualized environment is an
important requirement. Virtualization- do not expose
location/mechanisms of underlying hardware
Autonomous system: Self managed systems that can
work independently with high level guidance from
humans
Layered Architecture
Introduction Traditional Environment
Infrastructure

Layered
Architecture 11111111111111111111111
00000000000000000000000
Application 1 Application 2
00000000000000000000000
11111111111111111111111
Application n

Virtualization 00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
Further
00000000000000000000000
11111111111111111111111
Operating System

00000000000000000000000
11111111111111111111111
Reading

00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
Physical Hardware
Layered Architecture
Introduction Traditional Environment
Infrastructure

Layered
Architecture 11111111111111111111111
00000000000000000000000
Application 1
00000000000000000000000
11111111111111111111111
Application 2 Application n

Virtualization 00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
Further
00000000000000000000000
11111111111111111111111
Operating System

00000000000000000000000
11111111111111111111111
Reading

00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
Physical Hardware

Multiple OS

Application Application Application Application

1111111111
0000000000
0000000000
1111111111 0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
Operating System 0000000000
1111111111
Operating System
0000000000
1111111111
00000000000000000000000
11111111111111111111111 0000000000
1111111111
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
Physical Hardware
Layered Architecture
Introduction Traditional Environment
Infrastructure

Layered
Architecture 11111111111111111111111
00000000000000000000000
Application 1
00000000000000000000000
11111111111111111111111
Application 2 Application n

Virtualization 00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
Further
00000000000000000000000
11111111111111111111111
Operating System

00000000000000000000000
11111111111111111111111
Reading

00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
Physical Hardware

Multiple OS-not possible

Application Application Application Application

Infrastructure

Layered Multiple OS Hardware Problem

Architecture

Virtualization
OS caters to a specific hardware which is expressed in
Further terms of the Instruction Set Architecture (ISA)
Reading
ISA defines the instruction set for the processor,
registers, memory, and interrupt management
Specific OS is written for a specific ISA which in turn
abstracts a specific hardware
Further Applications are written for a specific OS
Running multiple OS on the same hardware
simultaneously is not possible
Layered Architecture
Introduction

Infrastructure

Layered Multiple OS Hardware Problem

Architecture

What about dual boot??

Storage is partitioned for each OS
Virtualization
Introduction

Infrastructure

Layered VM VM VM Virtual Machines

Architecture

Virtualization Layer of
Vitrual Machine Manager (VMM)
Further
11111111111111111111111
00000000000000000000000 Indirection
00000000000000000000000
11111111111111111111111
Hypervisor
Reading
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
Operating System

00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
Physical Hardware

Virtual machine (VM) can have different OS

Virtualization
Introduction

Infrastructure

Layered VM VM VM Virtual Machines

Architecture

Virtual machine (VM) can have different OS

VMs are one instance of this trend. With a VM, guest
OS and applications run in an environment that appears
and that behaves toward them as native hardware
would do, but also protects,manages, and limits them.
Virtualization
Introduction

Infrastructure

Layered
Architecture

Virtualization
Virtual machine implementations involve several
Further
components:
Reading At the base is the host, the underlying hardware system
that runs the virtual machines.
The virtual machine manager (VMM) (also known as a
hypervisor) creates and runs virtual machines by
providing an interface that is identical to the host
Each guest process (OS) is provided with a virtual copy
of the host
A single physical machine can thus run multiple
operating systems concurrently, each in its own virtual
machine.
Virtualization
Introduction

Infrastructure Virtualization technology is one of the fundamental

Layered components of cloud computing
Architecture

Virtualization
Virtualization is a broad concept that refers to the
Further creation of a virtual version hardware, a software
Reading
environment, storage, or a network.
It first came about in 1972 on IBM mainframes as a
method for multiple users to run tasks concurrently.
IBM VM /370 divided a mainframe into multiple virtual
machines, each running its own operating system
Minidisk was virtual disk drive allocated to each VM
Virtualization is to abstract the hardware of a single computer
(the CPU, memory, disk drives, network interface cards,
and so forth) into several different execution environments,
thereby creating the illusion that each separate environment
is running on its own private computer
Features of Virtualization
Introduction

Infrastructure
1. Increased security: virtualization of the execution
Layered environment allows increased security
Architecture

Virtualization
The ability to control the execution of a guest in a
Further
completely transparent manner delivering a secure,
Reading controlled execution environment.
The virtual machine represents an emulated
environment in which the guest is executed.
All the operations of the guest are generally performed
against the virtual machine, which then translates and
applies them to the host.
This level of indirection allows the VMM to control and
filter the activity of the guest, thus preventing some
harmful operations from being performed
Sensitive information contained in host can be hidden
without need to install complex security policies
Features of Virtualization
Introduction

Infrastructure Increased security is a requirement when dealing with

Layered untrusted code
Architecture
Applets downloaded from the Internet run in a
Virtualization
sandboxed Java Virtual Machine (JVM)
Further
Reading Providing them with limited access to the hosting
operating system resources.
Sandbox refers to an isolated execution environment where
potentially harmful instructions can be filtered and blocked
before being translated and executed in the real execution
environment.
2. Sharing: virtualization allows the creation of a
separate computing environments within the host.
Makes it possible to fully exploit the capabilities which
would otherwise be underutilized
Reduce number of active servers in data center leading
to reduction in power consumption
Features of Virtualization
Introduction

Infrastructure

Layered
Architecture
3.Aggregation- opposite process of sharing.
Virtualization

Further
Reading
Features of Virtualization
Introduction

Infrastructure

Layered
Architecture
3.Aggregation- opposite process of sharing.
Virtualization
A group of separate hosts can be tied together and
Further
represented to guests as a single virtual host.
Reading E.g., cluster management software harnesses the
physical resources of a homogeneous group of
machines and represents them as a single resource.
4. Emulation: guest programs are executed within an
environment that is controlled by the virtualization layer,
which ultimately is a program
This allows for controlling and tuning the environment
that is exposed to guests.
Features of Virtualization
Introduction

Infrastructure

Infrastructure 5. Isolation: virtualization allows providing

Layered guests-whether they are operating systems,
Architecture
applications, or other entities-with a completely
Virtualization
separate environment, in which they areexecuted.
Further
Reading It allows multiple guests to run on the same host without
interfering with each other
It provides a separation between the host and the
guest. The virtual machine can filter the activity of the
guest and prevent harmful operations against the host
6.Portability is another advantage of virtualization,
especially for execution virtualization techniques.
Virtual machine instances are normally represented by
one or more files that can be easily transported with
respect to physical systems
Features of Virtualization
Introduction

Infrastructure 5. Isolation: virtualization allows providing

Infrastructure

Layered
Architecture
Virtualization also has downsides
Virtualization

Further
Performance degradation: Performance is definitely
Reading one of the major concerns in using virtualization
technology. Since virtualization interposes an
abstraction layer between the guest and the host, the
guest can experience increased latencies.
Inefficiency and degraded user experience:
Virtualization can sometime lead to an inefficient use of
the host. In particular, some of the specific features of
the host cannot be exposed by the abstraction layer
and then become inaccessible.
Disadvantages of Virtualization
Introduction

Infrastructure
Security holes and new threats Virtualization opens
Layered
the door to a new and unexpected form of phishing.
Architecture
The capability of emulating a host in a completely
Virtualization
transparent manner led the way to malicious programs
Further
Reading that are designed to extract sensitive information from
the guest.

Phishing is a malicious practice aimed at capturing sensitive

user information, such as usernames and passwords, by
recreating an environment identical in functionalities and
appearance to the one that manages this information

Examples of these kinds of malware

BluePill moves the execution of the installed OS within
a virtual machine
SubVirt infects the guest OS, and when the virtual
machine is rebooted, it gains control of the host
Server Consolidation
Introduction

Infrastructure Virtualization technologies gained interested due to

Layered
Architecture following phenomenon:
Virtualization The continuous need for additional capacity -storage or
Further
Reading
compute power call for expansion of data centers
This condition, along with hardware underutilization,
has led to the diffusion of a technique called server
consolidation
Server consolidation is a technique for aggregating
multiple services and applications originally deployed
on different servers on one physical server.
Server consolidation allows us to reduce the power
consumption of a data center and also resolve
hardware underutilization for which virtualization
technologies are fundamental
Virtual Machine
Introduction

Infrastructure

Layered
Architecture 1 The fundamental idea behind a virtual machine is to
Virtualization
abstract the hardware of a single computer into several
Further
Reading different execution environments
2 Thereby creating the illusion that each separate
environment is running on its own private computer
3 Virtual machines first appeared commercially on IBM
mainframes in 1972
4 IBM VM /370 divided a mainframe into multiple virtual
machines, each running its own operating system
5 Minidisk was virtual disk drive allocated to each virtual
machine
Virtual Machine
Introduction
The virtualization requirements called for:
Infrastructure
Fidelity: A VMM provides an environment for programs
Layered
Architecture that is essentially identical to the original machine
Virtualization Performance: Programs running within that
Further
Reading
environment show only minor performance decreases
Safety: The VMM is in complete control of system
resources
These were established by Goldberg and Popek in 1974 as
three properties have to be satisfied by VMM:
1 Equivalence: A guest running under the control of a
VMM should exhibit the same behavior as when it is
executed directly on the physical host
2 Resource control: The VMM should be in complete
control of virtualized resources
3 Efficiency: A statistically dominant fraction of the
machine instructions should be executed without
intervention from the VMM
Benefits and Features
Introduction

Infrastructure The host system is protected from the virtual machines,

Layered just as the virtual machines are protected from each
Architecture

Virtualization
other
Further VMM have ability to freeze, or suspend, a running
Reading
virtual machine
VMMs allow copies and snapshots to be made of the
guest. The copy can be used to create a new VM or to
move a VM from one machine to another with its
current state intact
A virtual machine system is a perfect vehicle for
operating-system research
A major advantage of virtual machines in production
data-center use is system consolidation. It involves
taking two or more separate systems and running them
in virtual machines on one system
Benefits and Features
Introduction

Infrastructure

Layered
Architecture
Live migration- moves a running VM from one physical
Virtualization
server to another without interrupting its operation or
Further
Reading active network connections
Templating in which one standard virtual machine
image, including an installed and configured guest
operating system and applications, is saved and used
as a source for multiple running VMs
Cloud computing is made possible by virtualization in
which resources such as CPU, memory, and I/O are
provided as services to customers using Internet
technologies
Further Reading
Introduction

Infrastructure

Layered
Architecture

Virtualization

Further
Reading

1 Buyya, Rajkumar, Christian Vecchiola, and S. Thamarai Selvi. Mastering cloud computing: foundations
and applications programming. Newnes, 2013
Introduction

Infrastructure

Layered
Architecture

Virtualization

Further
Reading

Thank you
Course on Cloud Computing

Virtualization II

Dr. Gitanjali Chandwani Manocha

Assistant Professor
Department of Electronics & Communication Engg
Thapar Institute of Engineering and Technology, Patiala
Virtual Machine
Introduction

Machine
Reference
Model

Further
1 The fundamental idea behind a virtual machine is to
Reading abstract the hardware of a single computer into several
different execution environments
2 Thereby creating the illusion that each separate
environment is running on its own private computer
3 Virtual machines first appeared commercially on IBM
mainframes in 1972
4 IBM VM /370 divided a mainframe into multiple virtual
machines, each running its own operating system
5 Minidisk was virtual disk drive allocated to each virtual
machine
Virtual Machine
Introduction
The virtualization requirements called for:
Machine
Reference Fidelity: A VMM provides an environment for programs
Model
that is essentially identical to the original machine
Further
Reading Performance: Programs running within that
environment show only minor performance decreases
Safety: The VMM is in complete control of system
resources
These were established by Goldberg and Popek in 1974 as
three properties have to be satisfied by VMM:
1 Equivalence: A guest running under the control of a
VMM should exhibit the same behavior as when it is
executed directly on the physical host
2 Resource control: The VMM should be in complete
control of virtualized resources
3 Efficiency: A statistically dominant fraction of the
machine instructions should be executed without
intervention from the VMM
Machine Reference Model
Introduction

Machine
Reference
Model

Further
Reading

Hardware is expressed in terms of Instruction Set Architecture

(ISA)

It defines the instruction set for the processor, registers,

memory, and interrupt management.
ISA is the interface between hardware and software
It is important to the operating system (OS) developer &
developers of applications that directly manage the
underlying hardware
Machine Reference Model
Introduction

Machine
Reference
Model

Further
Reading

Application Binary Interface (ABI) separates the OS layer from the

applications and libraries

It covers details such as low-level data types, alignment, call

conventions and defines a format for executable programs
This interface allows portability of applications and libraries
across operating systems that implement the same ABI
Machine Reference Model
Introduction

Machine
Reference
Model

Further
Reading

The highest level of abstraction is represented by the application

programming interface (API)

It which interfaces applications to libraries and/or the

underlying operating system
Machine Reference Model
Introduction

Machine
Reference
Model

Further
Reading For any operation to be performed in the application
level API, ABI and ISA are responsible for making it
happen.
The high-level abstraction is converted into
machine-level instructions to perform the actual
operations supported by the processor
This layered approach simplifies the development and
implementation of computing systems, multitasking and
the coexistence of multiple executing environments
Machine Reference Model
Introduction

Machine
Reference The instruction set exposed by the hardware has been
Model
divided into different security classes that define who
Further
Reading can operate with them.
Nonprivileged instructions are those instructions that
can be used without interfering with other tasks
because they do not access shared resources.
E.g., all floating, fixed-point, and arithmetic instructions
Privileged instructions are executed under specific
restrictions & are mostly used for sensitive
operations, which expose (behavior-sensitive) or
modify (control-sensitive) the privileged state.
Behavior-sensitive instructions are those that operate
on the I/O
Control-sensitive instructions alter the state of the CPU
registers.
Security Rings and Privilege
Introduction

Machine
Modes
Reference
Model Some types of architecture
Further feature more than one class
Reading
of privileged instructions
A possible implementation
features a hierarchy of
privileges in the form of
ring-based security
Ring 0 is in the most privileged level used by the kernel
of the OS
Rings 1 and 2 are used by the OS-level services
Ring 3 in the least privileged level used by user
All the current systems support at least two different
execution modes: supervisor mode (Ring 0) and user
mode (Ring 3)
Execution Modes
Introduction

Machine
Reference
Supervisor mode
Model
All the instructions (privileged and nonprivileged) can be
Further
Reading executed without any restriction
This mode, also called master mode or kernel mode
It is generally used by the OS (or the hypervisor) to perform
sensitive operations on hardware-level resources

User mode
There are restrictions to control the machine-level resources
If code running in user mode invokes the privileged
instructions, hardware interrupts occur and trap the
potentially harmful execution of the instruction
There might be some instructions that can be invoked as
privileged instructions under some conditions and as
nonprivileged instructions under other conditions
Trap and Emulate
Introduction Implementation of virtual machine on dual-mode systems, where
Machine the underlying machine has only user mode and supervisory
Reference
Model (kernel) mode
Further
Reading

Virtual CPU (VCPU) it represents the state of the CPU as

views by guest machine
Trap is a software-generated interrupt caused by an error or
by a specific request from a user program that an OS service
be performed by executing a special operation called a
system call
Binary Translation
Introduction

Machine
Reference
Model

Further
Reading
Hypervisor
Introduction
The distinction between user and supervisor mode allows us
Machine
Reference to understand the role of the hypervisor
Model
The hypervisor runs above the supervisor mode, and from
Further
Reading here the prefix hyper- is used
In reality, hypervisors are run in supervisor mode, and the
division between privileged and nonprivileged instructions
has posed challenges in designing VMM.
It is expected that all the sensitive instructions will be
executed in privileged mode, which requires supervisor
mode in order to avoid traps.
Without this assumption it is impossible to fully emulate and
manage the status of the CPU for guest operating systems.
Unfortunately, this is not true for the original ISA, which
allows 17 sensitive instructions to be called in user mode.
This prevents multiple operating systems managed by a
single hypervisor to be isolated from each other, since
they are able to access the privileged state of the
processor and change it
Hypervisor
Introduction

Machine
Reference
Model

Further It is expected that in a hypervisor-managed

Reading
environment, all the guest operating system code will
be run in user mode in order to prevent it from directly
accessing the status of the CPU.
If there are sensitive instructions that can be called in
user mode (that is, implemented as nonprivileged
instructions), it is no longer possible to completely
isolate the guest OS
More recent implementations of ISA have solved this
problem by redesigning such instructions as privileged
ones
Terminologies
Introduction

Machine
Reference
Model

Further
Reading

Guest represents the system component that interacts

with the virtualization layer rather than with the host, as
would normally happen
Guest OS run on top of VMM along with applications
The host represents the original environment where the
guest is supposed to be managed
The virtualization layer is responsible for recreating the
same or a different environment where the guest will
operate
Virtual Machine
Introduction

Machine
Reference
Model

Machine The host system is protected from the virtual machines,

Reference
Model just as the virtual machines are protected from each
Further other
Reading
VMM have ability to freeze, or suspend, a running
virtual machine
VMMs allow copies and snapshots to be made of the
guest. The copy can be used to create a new VM or to
move a VM from one machine to another with its
current state intact
A virtual machine system is a perfect vehicle for
operating-system research
A major advantage of virtual machines in production
data-center use is system consolidation. It involves
taking two or more separate systems and running them
in virtual machines on one system
Benefits and Features
Introduction

Machine
Reference
Model

Further Live migration- moves a running VM from one physical

Reading
server to another without interrupting its operation or
active network connections
Templating in which one standard virtual machine
image, including an installed and configured guest
operating system and applications, is saved and used
as a source for multiple running VMs
Cloud computing is made possible by virtualization in
which resources such as CPU, memory, and I/O are
provided as services to customers using Internet
technologies
Further Reading
Introduction

Machine
Reference
Model

Further
Reading

1 Kai Hwang , Jack Dongarra , Geoffrey C. Fox Distributed and Cloud Computing: From Parallel Processing
to the Internet of Things. Morgan Kauffman 2011
2 Buyya, Rajkumar, Christian Vecchiola, and S. Thamarai Selvi. Mastering cloud computing: foundations
and applications programming. Newnes, 2013
3 Anthony Velte, Toby Velte and Robert Elsenpeter , Cloud Computing: A practical Approach Tata
McGrawHill, 2010, Second Edition
4 Judith Hurwitz, Robin Bllor, Marcia Kaufmann, Fern Halper, Cloud cOmputing for Dummies, 2009, Third
Edition
Introduction

Machine
Reference
Model

Further
Reading

Thank you
Course on Cloud Computing

Resource Management

Dr. Gitanjali Chandwani Manocha

Assistant Professor
Department of Electronics & Communication Engg
Thapar Institute of Engineering and Technology, Patiala
Introduction
Resource
Management

Resource In a cloud computing environment there are two actors

Provisioning
playing an important: cloud providers and cloud users
Resource
Scheduling

Resource
User’s objectives Cloud Provider objectives
Monitoring
Maximize performance Maximize Revenue
Qos based
Resource
Provisioning
Minimize finishing time Maximize Resource
Further Minimize cost Utilization
Reading

Users Request for the resources and cloud provider

allocates these to user for utilization.
Resource- CPU, Storage,Network, Power,etc.

Resource management is process of allocating these

resources to meet performance objective
Resource Management
Resource
Management

Resource
Provisioning

Resource
Scheduling

Resource
Monitoring

Qos based
Resource
Provisioning

Further
Reading

Cloud workload is an abstraction of work of that instance or

set of instances that are going to be executed
Resource Management
Resource
Management

Resource Cloud Users submits request for workload execution in

Provisioning
the form of work-load details
Resource
Scheduling Broker (resource provisioner) finds the suitable
Resource
Monitoring
resource(s) for a given workload and determines the
Qos based feasibility of provisioning of resources based on QoS
Resource
Provisioning requirements
Further Other responsibilities of broker include: release of extra
Reading
resources to resource pool, contains information of
provisioned resources and monitor performance to add
or remove resources
Broker sends requests to resource scheduler for
scheduling after successful provisioning of resources
After resource provisioning, resource scheduling is
done in second stage
Resource Management
Resource
Management

Resource
Provisioning

Resource All the provisioned resources are kept in resource

Scheduling
queue while other remaining resources are in resource
Resource
Monitoring pool
Qos based
Resource
Submitted workloads are processed in work-load
Provisioning queue. In this stage, scheduling agent maps the
Further
Reading
provisioned resources to given workload(s), execute
the workload(s) and release the resources back to
resources pool after successful completion of
work-load(s).
Based on QoS requirements, scheduling of resources
for adequate workloads is a challenging issue.
Resource Management
Resource
Management 1. Resource Provisioning is to identify adequate
Resource virtulized resources for a given workload based on QoS
Provisioning
requirements described by cloud user
Resource
Scheduling Cloud provider accepts the user request for resources
Resource and creates appropriate number of virtual machines
Monitoring
and allocates them to users as per the demand
Qos based
Resource Resource provisioning is also responsible to meet the
Provisioning user’s need based upon the quality of service (QoS),
Further
Reading
service level agreements (SLA) negotiations and match
the resources to the upcoming workloads.
2. Resource Scheduling is mapping and execution of
user workloads based on selected resources through
resource provisioning

Objective of resource provisioning with scheduling is to

provision the virtual machines to users without violation of
SLA and fulfill the users demand
Need of resource provisioning
Resource
Management

Resource
Provisioning Objectives: distributing the workload on resources and
Resource increasing their resource consumption but reducing the time
Scheduling

Resource
of workload execution
Monitoring
To detect and select the best resources for users
Qos based
Resource based upon their request
Provisioning

Further
Optimal resource
Reading
Resources needed to serve the user should be
minimum to maintain a desirable level of service quality
(minimum execution time and maximum throughput)
User gets the services in minimum cost and time while
service provider get the maximum profit without
affecting the violation of SLA
Resource Provisioning
Resource
Management Technique
Resource
Provisioning 1 On-demand provisioning
Resource
Scheduling
2 Advanced reservation
Resource 3 Spot instances
Monitoring

Qos based
Resource On-demand provisioning
Provisioning

Further
Provides resources quickly to urgent work-loads
Reading
Executing too many workloads on a single resource may
cause problem of interference which would leads to
performance degradation and over provisioning
Users to pay based upon the resources being used
If the demand exceeds the reserved value then additional
resources are provisioned
Generally on demand additional resources are allocated to
the users at higher cost than advanced reservation
resources
Provisioning Techniques
Resource
Management Advanced reservation
Resource Users to reserve the resources in advance for a specific time
Provisioning

Resource Used in federated cloud as well as elastic compute cloud

Scheduling
(EC2)
Resource
Monitoring Drawbacks: prediction of future demand and prices of cloud
Qos based resources is a challenging issue in this scheme
Resource
Provisioning Over- provisioning and under-provisioning type of problem
Further also occur in this technique
Reading

Spot instances
Allows customers to bid on unused resources at a
much lower cost
Major cloud service providers (AWS, Google, and
Azure) provide environment to use this scheme
Resources price rate vary frequently in spot instances
based on supply and demand
Resource Scheduling
Resource
Management

Resource
Provisioning
Scheduling is the way to determine, which activity should be
Resource
performed based upon the required quality of service (QoS)
Scheduling
parameters
Resource
Monitoring

Qos based
Resource scheduling comprises of two functions
Resource
Provisioning
1) Resource Allocation- allocate appropriate resources
Further
to the suitable workloads on time, so that applications
Reading can utilize the resources effectively
The amount of resources should be minimum for a
workload to maintain a desirable level of service quality,
or maximize throughput of a workload
2) Resource Mapping- mapping of workloads to
appropriate resources based on the QoS requirements
as specified by user in terms of SLA to minimize the
cost and execution time and maximize the profit
Static Scheduling
Resource
Management Scheduling algorithm can be categorized in two parts: static
Resource
Provisioning
and dynamic scheduling
Resource Static scheduling algorithms need the information
Scheduling

Resource
about the task and resource in advance
Monitoring
Static algorithms work well when variation in workload
Qos based
Resource is very less and behavior of the system is not varying
Provisioning
frequently
Further
Reading Thus, static algorithms are not a suitable choice for
cloud computing
They easy to implement but these algorithms don’t
optimize the QoS parameters
It doesn’t provide the good performance in real
environment. Example of static algorithms are first in
first out (FIFO),round robin (RR), shortest job first (SJF)
etc.
Dynamic Scheduling
Resource
Management

Resource There is no need of advance information about the jobs

Provisioning
and resources
Resource
Scheduling These algorithms are more efficient and accurate for
Resource
Monitoring
cloud environment because if any node is in overloaded
Qos based condition then they transfers the task from overloaded
Resource
Provisioning node to under loaded node
Further Dynamic algorithm needs to monitor the node
Reading
continuously
It is not easy to implement; complexity is high
E.g., dynamic round robin, heterogeneous earliest
finish time (HEFT), clustering based heterogeneous
with dupli- cation (CBHD), weighted least connection
(WLC), particle swarm optimization (PSO), ant colony
optimization (ACO)
online and offline mode
Resource
Management scheduling
Resource
Provisioning

Resource
Online
Scheduling
In on-line mode, a customer request is mapped with the
Resource
Monitoring running virtual machines when scheduler gets the
Qos based request from customer side
Resource
Provisioning Each task is scheduled only once, the scheduling result
Further
Reading
remains unchanged

Offline
Offline scheduling is called batch mode scheduling in
which upcoming application request is allocated to
resources based upon predefined moments
It is used to calculate the processing time of larger
number of tasks
Preemptive and non-preemptive
Resource
Management scheduling
Resource
Provisioning

Resource
Scheduling

Resource
Monitoring
Preemptive
Qos based The tasks can be interrupted at the current execution
Resource
Provisioning and can be migrated to other resources
Further
Reading
Non-preemptive
When a task is allocated to cloud resource, it will not be
free until the task is finished i.e. task is executed
completely at the resource without being interrupted
Resource Monitoring
Resource
Management
Performance optimization can be best achieved by an
Resource
Provisioning efficient monitoring of the utilization of computing
Resource resources
Scheduling

Resource In SLA, both the parties (cloud provider and cloud user)
Monitoring
should have specified the possible deviations to
Qos based
Resource achieve appropriate quality attributes
Provisioning

Further
The resource monitoring system collects the resource
Reading usages by measuring through performance metrics
such as CPU and memory utilization
Cloud provider needs to retain the adequate number of
resources to deliver the continuous service to cloud
consumer during peak load
Resource monitoring is used to take care of important
QoS requirements like security, availability,
performance, etc. during workload execution.
Resource monitoring
Resource
Management

Resource
Provisioning

Resource
Scheduling
There are two main aspects of resource monitoring
Resource
Monitoring (i) Consumer wants to execute their workload at minimum
Qos based
Resource cost and minimum time without violation of SLA
Provisioning
(ii) Provider wants to execute the workload with minimum
Further
Reading number of resources
Resource monitoring is a vital part of resource
management to measure the SLA deviation, QoS
requirements and resource usages
QoS based Resource
Resource
Management Provisioning
Resource
Provisioning
Analyze the workloads, categorize them on the basis of
Resource
common patterns and then provision the cloud workloads
Scheduling before actual scheduling
Resource QoS metric based resource provisioning technique is
Monitoring
efficient in reducing execution time and execution cost
Qos based
Resource
Provisioning

Further
Reading
Further Reading
Resource
Management

Resource
Provisioning

Resource
Scheduling

Resource
Monitoring
1 Singh, Sukhpal, and Inderveer Chana. "A survey on resource scheduling in cloud computing: Issues and
Qos based challenges." Journal of grid computing 14.2 (2016): 217-264.
Resource
Provisioning 2 Singh, Sukhpal, and Inderveer Chana. "Cloud resource provisioning: survey, status and future research
directions." Knowledge and Information Systems 49.3 (2016): 1005-1069.
Further 3 Kumar, Mohit, et al. "A comprehensive survey for scheduling techniques in cloud computing." Journal of
Reading Network and Computer Applications 143 (2019): 1-33.
Resource
Management

Resource
Provisioning

Resource
Scheduling

Resource
Monitoring

Qos based
Resource
Provisioning
Thank you
Further
Reading
Course on Cloud Computing

Economic Benefits of Cloud Computing

Dr. Gitanjali Chandwani Manocha

Assistant Professor
Department of Electronics & Communication Engg
Thapar Institute of Engineering and Technology, Patiala
Economic Benefits
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics
Utility Pricing
The main drivers of cloud computing are economy of scale
Further and simplicity of software delivery and its operation.
Reading
Cloud computing allows:
Reducing the capital costs associated to the IT
infrastructure
Eliminating the depreciation or lifetime costs associated
with IT capital assets
Replacing software licensing with subscriptions Cutting
the maintenance and administrative costs of IT
resources
Direct Cost
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models The biggest benefit of Cloud is financial
Characteristics
Utility Pricing Pay-as-you-go model offered by cloud providers
Further
Reading
Eliminates/reduces capital cost: cost occurred in
purchasing an asset that is useful in the production of
goods or the rendering of services
Budget spent on IT infrastructure and software
constituted a significant expense for enterprises.
Owning a small or medium-sized datacenter introduces
several operational costs in terms of maintenance,
electricity, and cooling.
Additional operational costs are occurred in maintaining
an IT department and an IT support center.
other costs are triggered by the purchase of potentially
expensive software.
Direct Cost
Economic
Benefits
Direct Cost Cloud computing model shifts the capital costs into
Indirect Cost
Pricing Models operational costs inducted by renting the
Characteristics
Utility Pricing infrastructure and paying subscriptions for the use of
Further
Reading
software.
These costs can be controlled according to needs
Cloud computing also introduces reductions in
administrative and maintenance costs.
Depreciation costs disappears: IT needs are served by
the cloud There are no IT capital assets that depreciate
over time
Cloud computing, especially IaaS-based solutions, help
manage unplanned capital costs generated by the
short term needs
These costs can be turned into operational costs that
last as long as there is a need for them.
Indirect Cost
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics
Utility Pricing

Further
Reading
Cloud computing eliminates some indirect costs that are
generated by IT assets
Software licensing and support- use software
applications on a subscription basis thus there is no
need for licensing fee
Carbon footprint emissions- Datacenter consolidation
results in smaller carbon footprint
Pricing Models
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics
Utility Pricing There are three different pricing strategies that are adopted
Further by the cloud providers
Reading
Tiered pricing: In this model, cloud services are offered
in several tiers, each of which offers a fixed computing
specification and SLA at a specific price per unit of
time.
This model is used by Amazon for pricing the EC2
service, which makes available different server
configurations in terms of computing capacity (CPU
type and speed, memory) that have different costs per
hour.
Pricing Models
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics Per-unit pricing. This model is more suitable to cases
Utility Pricing
where the principal source of revenue for the cloud
Further
Reading provider is determined in terms of units of specific
services, such as data transfer and memory allocation.
This model is used, for example, by GoGrid, which
makes customers pay according to RAM/hour units for
the servers deployed in the GoGrid cloud.
Subscription-based pricing. This is the model used
mostly by SaaS providers in which users pay a periodic
subscription fee for use of the software or the specific
component services that are integrated in their
applications.
Economic Chracteristics
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models Common infrastructure: resources are pooled and
Characteristics
Utility Pricing dynamically shared infrastructure
Further
Reading Location independence: ubiquitous and responsive
service. It enhances the user experience and
convenience; dispersion also enhances availability.
Online accessibility: Without networks, there is no
cloud, key enablers as services are accessed over a
network
Utility pricing: creating value and with usage-sensitive
pricing charges for resources are all exactly aligned
On-Demand resource: right quantity of the right
resources available exactly and only when needed.
Is cloud less expensive
Economic
Benefits
It is cloud computing, not cheap computing. Cost benefit is
Direct Cost
Indirect Cost
unique to each customer.
Pricing Models
Characteristics
It is possible that the cloud delivers resources and
Utility Pricing
services at a lower unit cost than an enterprise
Further
Reading Also possible that cloud costs more on unit-cost basis
Unit cost is cost for server-hours or gigabyte-months
Whether to use cloud or not?
Calls for characterization of Unit cost of an on-demand,
pay-per-use cloud resource relative to the unit cost of a
dedicated resource
Dedicated resource may be owned and incur a cost for
depreciation
May be leased and incur a monthly fee
May be financed, and represent an ongoing principal
plus interest fee to the bank
Utility Pricing
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics
Utility Pricing
Utility premium U is ratio of Cloud unit cost (C) to
Further
Reading Baseline (owned) unit cost (B)
C
U= (1)
B
If U =1 the cloud costs the same as ownership
If U<1, the cloud is cheaper
U may vary over time but is assumed to be constant in
this framework
Utility Pricing
Economic
Benefits Peak of D(t) is peak demand:P
Direct Cost
Indirect Cost
Pricing Models
Average of D(t) over period of 0 to T is Average
Characteristics
Utility Pricing
demand: A
Further
Reading
Utility Pricing
Economic
Benefits
Direct Cost
CT Cloud cost for time T can be expresses as
Indirect Cost
Pricing Models
Z T
Characteristics
Utility Pricing
CT = U × B × D(t) = U × B × A × T (2)
0
Further
Reading
Utility Pricing
Economic
Benefits
Direct Cost
CT Cloud cost for time T can be expresses as
Indirect Cost
Pricing Models
Z T
Characteristics
Utility Pricing
CT = U × B × D(t) = U × B × A × T (2)
0
Further
Reading
BT Baseline cost for time T can be expresses as

BT = B × T × P (3)

On comparing these equations we find that cloud will

be economically beneficial when

CT < BT U ×B×A×T < B×T ×P (4)

P
U< (5)
A
Utility Premium is less that the ratio of peak demand to
average demand
Utility Pricing
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics If U< 1 then cloud is always a good idea, regardless of
Utility Pricing

Further
the demand profile as A≤P by definition.
Reading
Utility Pricing
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics If U< 1 then cloud is always a good idea, regardless of
Utility Pricing

Further
the demand profile as A≤P by definition.
Reading
If clouds cost the same i.e., U=1 then relative total cost
then depends on whether
Utility Pricing
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics If U< 1 then cloud is always a good idea, regardless of
Utility Pricing

Further
the demand profile as A≤P by definition.
Reading
If clouds cost the same i.e., U=1 then relative total cost
then depends on whether
A=P (demand is flat): no pay-per-use-go benefit
Utility Pricing
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics If U< 1 then cloud is always a good idea, regardless of
Utility Pricing

Further
the demand profile as A≤P by definition.
Reading
If clouds cost the same i.e., U=1 then relative total cost
then depends on whether
A=P (demand is flat): no pay-per-use-go benefit
A<P (there is some variation)
Utility Pricing
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics If U< 1 then cloud is always a good idea, regardless of
Utility Pricing

Further
the demand profile as A≤P by definition.
Reading
If clouds cost the same i.e., U=1 then relative total cost
then depends on whether
A=P (demand is flat): no pay-per-use-go benefit
A<P (there is some variation)
If U>1,cloud is more expensive on a unit-cost basis
Utility Pricing
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics If U< 1 then cloud is always a good idea, regardless of
Utility Pricing

Further
the demand profile as A≤P by definition.
Reading
If clouds cost the same i.e., U=1 then relative total cost
then depends on whether
A=P (demand is flat): no pay-per-use-go benefit
A<P (there is some variation)
If U>1,cloud is more expensive on a unit-cost basis
A=P (demand is flat): owning resources would be less
expensive
Utility Pricing
Economic
Benefits
Direct Cost
Indirect Cost
Pricing Models
Characteristics If U< 1 then cloud is always a good idea, regardless of
Utility Pricing

Further
the demand profile as A≤P by definition.
Reading
If clouds cost the same i.e., U=1 then relative total cost
then depends on whether
A=P (demand is flat): no pay-per-use-go benefit
A<P (there is some variation)
If U>1,cloud is more expensive on a unit-cost basis
A=P (demand is flat): owning resources would be less
expensive
A<P (there is some variation): if demand is spikier
compared utility premium U that is P/A >U the cloud still
can be cheaper
Further Reading
Economic
Benefits

Further
Reading

1 Kai Hwang , Jack Dongarra , Geoffrey C. Fox Distributed and Cloud Computing: From Parallel Processing
to the Internet of Things. Morgan Kauffman 2011
2 Rajkumar Buyya, James Broberg and Goscinski Author Name, Cloud Computing Principles and
Paradigms, John Wiley and Sons 2012, Second Edition
3 Gerard Blokdijk, Ivanka Menken,The Complete Cornerstone Guide to Cloud Computing Best Practices,
Emereo Pvt Ltd, 2009, Second Edition
4 Anthony Velte, Toby Velte and Robert Elsenpeter , Cloud Computing: A practical Approach Tata
McGrawHill, 2010, Second Edition
5 Judith Hurwitz, Robin Bllor, Marcia Kaufmann, Fern Halper, Cloud cOmputing for Dummies, 2009, Third
Edition
Economic
Benefits

Further
Reading

Thank you
Course on Cloud Computing

Cloud Services and platforms I

Dr. Gitanjali Chandwani Manocha

Assistant Professor
Department of Electronics & Communication Engg
Thapar Institute of Engineering and Technology, Patiala
Cloud Reference Model
Cloud
Reference
Model

Compute
Services

Storage
Services

Application
Service

Queuing
Service

Database
Services

Content
Delivery
Service

Further
Reading

Courtesy[1]
Cloud Reference Model
Cloud
Reference
Model

Compute
Services

Storage
Services

Application Infrastructure &

Service
Facilities Layer
Queuing
Service Includes the physical
Database infrastructure such
Services
as datacenter
Content
Delivery facilities, electrical
Service
and mechanical
Further equipment, etc
Reading
Cloud Reference Model
Cloud
Reference
Model

Compute
Services

Storage
Services

Application
Service

Queuing Hardware Layer

Service

Database
Includes physical
Services compute, network
Content and storage
Delivery
Service hardware
Further
Reading
Cloud Reference Model
Cloud
Reference
Model

Compute
Services

Storage
Services

Application Virtualization Layer

Service

Queuing
Partitions the
Service physical hardware
Database resources into
Services
multiple virtual
Content
Delivery resources that
Service
enabling pooling of
Further
Reading
resources
Cloud Reference Model
Cloud
Reference
Model

Compute
Services

Storage Platform & Middleware Layer

Services

Application
Builds upon the IaaS layers
Service below and provides
Queuing standardized stacks of
Service
services such as database
Database
Services service, queuing service,
Content application frameworks and
Delivery run-time environments,
Service

Further
messaging services,
Reading monitoring services,
analytics services, etc
Cloud Reference Model
Cloud
Reference
Model

Compute
Services

Storage
Services

Application
Service Service Management
Queuing Layer
Service

Database
Provides APIs for
Services requesting,
Content managing and
Delivery
Service monitoring cloud
Further resources
Reading
Cloud Reference Model
Cloud
Reference
Model

Compute
Services

Storage
Services Applications Layer
Application
Service
Includes SaaS such
Queuing
as Email, cloud
Service storage application,
Database productivity
Services
applications,
Content
Delivery management portals,
Service
customer
Further
Reading
self-service portals,
etc
Compute Services
Cloud
Reference Compute services provide dynamically scalable
Model

Compute
compute capacity in the cloud.
Services
Amazon EC2
Compute resources can be provisioned on-demand in
Storage the form of virtual machines. Virtual machines can be
Services
created from standard images provided by the cloud
Application
Service service provider or custom images created by the
Queuing users.
Service

Database
Compute services can be accessed from the web
Services
consoles of these services that provide graphical user
Content
Delivery interfaces for provisioning, managing and monitoring
Service
these services.
Further
Reading Cloud service providers also provide APIs for various
programming languages that allow developers to
access and manage these services programmatically
E.g., Amazon Elastic Compute Cloud, Google Compute
Engine, Windows Azure Virtual Machine
Compute Services - Amazon
Cloud
Reference
Model
EC2
Compute
Services Amazon Elastic Compute Cloud (EC2) is a compute
Amazon EC2

Storage
service (Iaas) provided by Amazon
Services
Virtual computing environments is known as instances
Application
Service Preconfigured templates for your instances, known as
Queuing
Service
Amazon Machine Images (AMIs), that package the bits
Database you need for your server (including the operating
Services
system and additional software
Content
Delivery Instances are created using AMI as templates, which
Service

Further
are specialized by selecting the number of cores, their
Reading computing power, and the installed memory
Various configurations of CPU, memory, storage, and
networking capacity for your instances, known as
instance types
Compute Services - Amazon
Cloud
Reference
Model
EC2
Compute
Services
Amazon EC2
Currently Available Instances of EC2:
Storage https://aws.amazon.com/ec2/instance-types/
Services
Launching EC2 Instances
Application
Service To launch a new instance click on the launch instance
Queuing
Service
button
Database This will open a wizard where user can select the
Services
Amazon machine image (AMI) with which you want to
Content
Delivery launch the instance
Service

Further User can also create their own AMIs with custom
Reading
applications, libraries and data.
Instances can be launched with a variety of operating
systems.
Storage Services
Cloud
Reference
Model

Compute
Services
Cloud storage services allow storage and retrieval of
Storage any amount of data, at any time from anywhere on the
Services
Amazon S3
web.
Amazon elastic block
store Most cloud storage services organize data into
Application
Service
buckets or containers.
Queuing Scalability - Cloud storage services provide high
Service

Database
capacity and scalability. Objects upto several tera-bytes
Services in size can be uploaded and multiple
Content
Delivery
buckets/containers can be created on cloud storages.
Service
Replication- When an object is uploaded it is replicated
Further
Reading at multiple facilities and/or on multiple devices within
each facility.
Storage Services
Cloud
Reference
Model
Access Policies -Cloud storage services provide several
Compute
Services security features such as Access Control Lists (ACLs),
Storage bucket/container level policies, etc. ACLs can be used
Services
Amazon S3 to selectively grant access permissions on individual
Amazon elastic block
store
objects. Bucket/container level policies can also be
Application
Service defined to allow or deny permissions across some or all
Queuing of the objects within a single bucket/container.
Service

Database
Encryption - Cloud storage services provide Server
Services Side Encryption (SSE) options to encrypt all data
Content
Delivery
stored in the cloud storage.
Service
Consistency - Strong data consistency is provided for
Further
Reading all upload and delete operations. Therefore, any object
that is uploaded can be immediately downloaded after
the upload is complete
Amazon Simple Storage
Cloud
Reference
Model
Service(S3)
Compute
Services

Storage
Services
It is an online cloud-based data storage infrastructure
Amazon S3
Amazon elastic block
for storing and retrieving any amount of data.
store

Application
S3 provides highly reliable, scalable, fast, fully
Service redundant and affordable storage infrastructure.
Queuing
Service Data stored on S3 is organized in the form of buckets
Database
Services
Uploading Files to Buckets
Content
S3 console provides simple wizards for creating a new
Delivery bucket and uploading files.
Service
You can upload any kind of file to S3.
Further
Reading While uploading a file, you can specify the redundancy
and encryption options and access permissions.
Amazon S3
Cloud
Reference Two core components of S3 are:
Model Buckets: They represent virtual containers in which to
Compute store object
Services
Objects represent the content that is actually stored.
Storage
Services Objects can also be enriched with metadata that can be
Amazon S3
Amazon elastic block
used to tag the stored content with additional
store
information
Application
Service S3 has been designed to provide a simple storage
Queuing service that’s accessible through a Representational
Service
State Transfer (REST) interface.
Database
Services Access to S3 is provided with RESTful Web services.
Content These express all the operations that can be performed
Delivery
Service
on the storage in the form of HTTP requests (GET, PUT,
Further
DELETE, HEAD, and POST)
Reading PUT/POST requests add new content to the store
GET/HEAD requests are used to retrieve content and
information
DELETE requests are used to remove elements or
information attached to them.
S3 Key Concepts
Cloud
Reference
Model

Compute
1 The storage is organized in a two-level hierarchy. S3
Services organizes its storage space into buckets that cannot be
Storage further partitioned.
Services
Amazon S3 This means that it is not possible to create directories or
Amazon elastic block
store other kinds of physical groupings for objects stored in a
Application bucket
Service

Queuing
2 Stored objects cannot be manipulated like standard
Service files
Database
Services
S3 has been designed to essentially provide storage for
Content
objects that will not change over time
Delivery it does not allow renaming, modifying, or relocating an
Service
object
Further
Reading Once an object has been added to a bucket, its content
and position is immutable, and the only way to change it
is to remove the object from the store and add it again
Amazon S3 Key Concepts
Cloud
Reference 3 Content is not immediately available to users
Model

Compute
The main design goal of S3 is to provide an eventually
Services consistent data store -because it is a large distributed
Storage storage facility
Services
Amazon S3
As a result changes are not immediately reflected
Amazon elastic block
store For instance, S3 uses replication to provide
Application redundancy and efficiently serve objects across the
Service
globe; this practice introduces latencies when adding
Queuing
Service
objects to the store-especially large ones-which are not
Database
available instantly across the entire globe.
Services 4 Requests will occasionally fail
Content
Delivery Due to the large distributed infrastructure being
Service managed, requests for object may occasionally fail
Further
Reading
Under certain conditions, S3 can decide to drop a
request by returning an internal server error
Therefore, it is expected to have a small failure rate
during day-to-day operations, which is generally not
identified as a persistent failure
Bucket
Cloud
Reference
Model
A bucket is a container of objects.
Compute
Services It can be thought of as a virtual drive hosted on the S3
Storage
Services
distributed storage, which provides users with a flat
Amazon S3
Amazon elastic block
store to which they can add objects
store

Application
Buckets are top-level elements of the S3 storage
Service architecture and do not support nesting. That is, it is
Queuing
Service
not possible to create "subbuckets" or other kinds of
Database
physical divisions.
Services
A bucket is located in a specific geographic location
Content
Delivery and eventually replicated for fault tolerance and better
Service

Further
content distribution
Reading
Users can select the location at which to create
buckets, which by default are created in Amazon’s U.S.
data centers.
Bucket
Cloud
Reference Once a bucket is created, all the objects that belong to
Model

Compute
the bucket will be stored in the same availability zone of
Services the bucket.
Storage
Services Users create a bucket by sending a PUT request to
Amazon S3
Amazon elastic block
http://s3.amazonaws.com/ with the name of the bucket
store
and, if they want to specify the availability zone,
Application
Service additional information about the preferred location
Queuing
Service
The content of a bucket can be listed by sending a GET
Database request specifying the name of the bucket
Services
Once created, the bucket cannot be renamed or
Content
Delivery relocated
Service

Further
If it is necessary to do so, the bucket needs to be
Reading deleted and recreated
The deletion of a bucket is performed by a DELETE
request, which can be successful if and only if the
bucket is empty
Objects
Cloud
Reference
Model

Compute
Objects: fundamental entities stored in Amazon S3
Services
Objects consist of object data and metadata
Storage
Services
Amazon S3
The data portion is opaque to Amazon S3
Amazon elastic block
store The metadata is a set of name-value pairs that
Application describe the object
Service

Queuing These include some default metadata, such as the date

Service
last modified, and standard HTTP metadata, such as
Database
Services Content-Type
Content
Delivery
One can also specify custom metadata at the time the
Service
object is stored.
Further
Reading An object is uniquely identified within a bucket by a key
(name) and a version ID. For more information, see
Keys and Using Versioning. Keys
Keys
Cloud
Reference
Model

Compute
Services
A key is the unique identifier for an object within a
Storage
bucket
Services
Amazon S3 Every object in a bucket has exactly one key
Amazon elastic block
store
The combination of a bucket, key, and version ID
Application
Service uniquely identify each object.
Queuing
Service
Every object in Amazon S3 can be uniquely addressed
Database
through the combination of the web service endpoint,
Services bucket name, key, and optionally, a version
Content
Delivery For example, in the URL
Service
https://doc.s3.amazonaws.com/2006-03-
Further
Reading 01/AmazonS3.wsdl, "doc" is the name of the bucket
and "2006-03-01/AmazonS3.wsdl" is the key.
Amazon Elastic Block Store
Cloud
Reference The Amazon Elastic Block Store (EBS) allows AWS users to
Model
provide EC2 instances with persistent storage in the form
Compute
Services
of volumes that can be mounted at instance startup
Storage They accommodate up to 1 TB of space and are accessed
Services
Amazon S3
through a block device interface, thus allowing users to
Amazon elastic block
store
format them according to the needs of the instance they are
Application
connected to (raw storage, file system, or other)
Service
The content of an EBS volume survives the instance life
Queuing
Service
cycle and is persisted into S3
Database EBS volumes can be cloned, used as boot partitions, and
Services constitute durable storage since they rely on S3 and it is
Content
Delivery
possible to take incremental snapshots of their content.
Service EBS volumes normally reside within the same availability
Further
Reading
zone of the EC2 instances that will use them to maximize the
I/O performance
The expense related to a volume comprises the cost
generated by the amount of storage occupied in S3 and by
the number of I/O requests performed against the volume.
Application Service
Cloud
Reference
Model

Compute Cloud application service services:

Services

Storage
Application Runtimes & Frameworks- Google
Services AppEngine, Window Azure Website
Application
Service Queuing service- Amazon SQS, Google task queue
Google AppEngine
service, Window Azure queue service
Queuing
Service Email service- Amazon Simple Email Service, Google
Database
Services
Email Service
Content notification service- Amazon Simple Notification
Delivery
Service Service, Google cloud messaging
Further
Reading media service-Amazon Elastic Transcoder, Google
Image Manipulation Service
Content Delivery Service- Amazon CloudFront
Application Runtimes &
Cloud
Reference
Model
Frameworks
Compute
Services

Storage
Services Cloud-based application runtimes and frameworks
Application
Service
allow developers to develop and host applications in the
Google AppEngine cloud
Queuing
Service Support for various programming languages:
Database Application runtimes provide support for programming
Services
languages (e.g., Java, Python, or Ruby).
Content
Delivery
Service
Resource Allocation: Application runtimes
Further automatically allocate resources for applications and
Reading
handle the application scaling, without the need to run
and maintain servers.
Google AppEngine
Cloud
Reference Google AppEngine (GAE) is a PaaS implementation
Model
that provides services for developing and hosting
Compute
Services scalable Web applications
Storage It is a distributed and scalable runtime environment that
Services

Application
leverages Google’s distributed infrastructure to scale
Service out applications facing a large number of requests by
Google AppEngine

Queuing
allocating more computing resources to them and
Service balancing the load among them
Database
Services
The runtime is completed by a collection of services
Content that allow developers to design and implement
Delivery
Service applications that naturally scale on AppEngine
Further Developers can develop applications in Java, Python,
Reading
and Go, a new programming language developed by
Google to simplify the development of Web applications
Application usage of Google resources and services is
metered by AppEngine, which bills users when their
applications finish their free quotas.
GAE- features
Cloud
Reference
Model

Compute
Runtimes: App Engine provides runtime environments
Services for Java, Python, PHP and Go programming language.
Storage
Services Sandbox: Applications run in a secure sandbox
Application environment isolated from other applications. The
Service
Google AppEngine sandbox environment provides a limited access to the
Queuing underlying operating system
Service

Database Datastore: provides a no-SQL data storage service

Services

Content
Authentication: App Engine applications can be
Delivery
Service
integrated with Google Accounts for user authentication
Further URL Fetch service: allows applications to access
Reading
resources on the Internet, such as web services
Other services: Email service, Image Manipulation
service, Memcache, Task Queues, Cron jobs
GAE- features
Cloud
Reference
Model

Compute
Services
MemCache: AppEngine provides caching services by
Storage
means of MemCache.
Services
This is a distributed in-memory cache that is optimized
Application
Service for fast access and provides developers with a volatile
Google AppEngine
store for the objects that are frequently accessed
Queuing
Service
The caching algorithm implemented by MemCache will
Database
automatically remove the objects that are rarely
Services accessed
Content The use of MemCache can significantly reduce the
Delivery
Service access time to data
Further Developers can structure their applications so that each
Reading
object is first looked up into MemCache and if there is a
miss, it will be retrieved from DataStore and put into the
cache for future lookups.
GAE- features
Cloud
Reference
Model

Compute
Services

Storage Task queues

Services
Task Queues allow applications to submit a task for a
Application
Service later execution.
Google AppEngine
This service is particularly useful for long computations
Queuing
Service
that cannot be completed within the maximum response
time of a request handler
Database
Services The service allows users to have up to 10 queues that
Content can execute tasks at a configurable rate.
Delivery
Service The queue is designed to reexecute the task in case of
Further
failure in order to avoid transient failures preventing the
Reading task from a successful completion.
GAE- features
Cloud
Reference
Model

Compute
Services
Cron jobs
Storage
Services It might be possible that the required operation needs to
Application be performed at a specific time of the day, which does
Service
Google AppEngine
not coincide with the time of the Web request.
Queuing It is possible to schedule such operation at the desired
Service time by using the Cron Jobs service.
Database This service operates similarly to Task Queues but
Services
invokes the request handler specified in the task at a
Content
Delivery given time and does not reexecute the task in case of
Service
failure
Further
Reading
This behavior can be useful to implement maintenance
operations or send periodic notifications
Amazon Simple Queue Service
Cloud
Reference
Model

Compute Amazon SQS constitutes disconnected model for

Services
exchanging messages between applications by means of
Storage
Services message queues, hosted within the AWS infrastructure.
Application
Service
It is a queuing service from Amazon
Queuing SQS is a distributed queue that supports short
Service
messages of up to 256 KB in size.
Database
Services
SQS supports multiple writers and readers and locks
Content
Delivery messages while they are being processed.
Service
High Availability
Further
Reading To ensure high availability SQS service trade-offs on
the first in, first out capability and does not guarantee
that messages will be delivered in FIFO order.
Applications that require FIFO ordering can place
sequencing information in each message so that they
can be re-ordered after retrieving from a queue.
Database Services
Cloud
Reference Cloud database services allow to set-up and operate
Model relational or non-relational databases in the cloud.
Compute Relational Databases E.g., MySQL, Oracle, SQL Server, etc.
Services

Storage
Non-relational Databases The non-relational (No-SQL)
Services databases provided by cloud service providers are mostly
Application proprietary solutions.
Service
Scalability Cloud database services allow provisioning as
Queuing
Service much compute and storage resources as required to meet
Database the application workload levels. Provisioned capacity can be
Services
scaled-up or down.
Content
Delivery
Reliability Cloud database services are reliable and provide
Service automated backup and snapshot options.
Further Performance Cloud database services provide guaranteed
Reading
performance with options such as guaranteed input/output
operations per second (IOPS) which can be provisioned
upfront.
Security Several security features are provided to restrict the
access to the database instances and stored data, such as
network firewalls and authentication mechanisms.
Amazon Relational Database
Cloud
Reference Amazon RDS is relational database service that relies
Model
on the EC2 infrastructure and is managed by Amazon.
Compute
Services Developers don’t have to worry about configuring the
Storage storage for high availability or designing failover
Services

Application
strategies
Service
Multi-AZ deployment: provides users with a failover
Queuing
Service infrastructure for their RDBMS solutions
Database The high-availability solution is implemented by keeping
Services in standby synchronized copies of the services in
Content
Delivery
different availability zones that are activated if the
Service primary service goes down.
Further
Reading Read replicas: provides users with increased
performance for applications that are heavily based on
database reads. In this case, Amazon deploys copies
of the primary service that are only available for
database reads, thus cutting down the response time of
the service.
Amazon SimpleDB
Cloud
Reference
Model

Compute
Services
Amazon SimpleDB is a lightweight, highly scalable, and
Storage flexible data storage solution for applica tions that do
Services
not require a fully relational model for their data
Application
Service SimpleDB provides support for semistructured data, the
Queuing
Service
model for which is based on the concept of domains,
Database
items, and attributes
Services
Compared to the relational model, this model provides
Content
Delivery fewer constraints on the structure of data entries, thus
Service

Further
obtaining improved performance in querying large
Reading quantities of data.
SimpleDB implements a relaxed constraint model,
which leads to eventually consistent data.
Content Delivery Service
Cloud
Reference
Model Cloud-based content delivery service include Content
Compute Delivery Networks (CDNs).
Services

Storage CDN is a distributed system of servers located across

Services
multiple geographic locations to serve content to end-
Application
Service users with high availability and high performance.
Queuing
Service
CDNs are useful for serving static content such as text,
Database images, scripts, etc., and streaming media.
Services

Content
CDNs have a number of edge locations deployed in
Delivery multiple locations, often over multiple backbones.
Service
CloudFront
Requests for static for streaming media content that is
Further
Reading served by a CDN are directed to the nearest edge
location.
E.g., Amazon CloudFront, Windows Azure Content
Delivery Network
CloudFront
Cloud
Reference CloudFront is an implementation of a CDN on top of the
Model Amazon distributed storage infrastructure
Compute
Services
It leverages a collection of edge servers strategically located
Storage
around the globe to better serve requests for static and
Services streaming Web content to reduce transfer time
Application The content can be static (HTTP and HTTPS) or streaming
Service
(Real Time Messaging Protocol, or RMTP)
Queuing
Service The origin server hosting the original copy of the dis tributed
Database content can be an S3 bucket, an EC2 instance, or a server
Services external to the Amazon network.
Content
Delivery
It is necessary to create a distribution: origin server, which
Service contains the original version of the content being distributed,
CloudFront
and it is referenced by a DNS domain under the
Further
Reading Cloudfront.net domain name
Once the distribution is created, it is sufficient to reference
the distribution name, and the CloudFront engine will redirect
the request to the closest replica and eventually download
the original version from the origin server if the content is not
found or expired on the selected edge server
Further Reading
Cloud
Reference
Model

Compute
Services

Storage
Services

Application
Service
1 Bahga, Arshdeep, and Vijay Madisetti. Cloud computing: A hands-on approach. CreateSpace
Queuing Independent Publishing Platform, 2013
Service
2 Buyya, Rajkumar, Christian Vecchiola, and S. Thamarai Selvi. Mastering cloud computing: foundations
Database and applications programming. Newnes, 2013
Services

Content
Delivery
Service

Dr. Gitanjali Chandwani Manocha

Assistant Professor
Department of Electronics & Communication Engg
Thapar Institute of Engineering and Technology, Patiala
Case Study 1
Case Study 1

Case Study 2

Case Study 3 Problem Statement

Case Study 4
Siyaram Autos (SRA) is a manufacturing company that
Case Study 5

Case Study 6
manufactures vehicle parts with a turn over 30-35
Case Study 7
lakhs/annum. The core competence of the company is to
Case Study 8 design and manufacture vehicle parts . The company does
Case Study 9 not have an IT department. The competitor company Guru
Further Dev Autos (GDA) uses off the shelf software. The market
Reading
competition forces SRA to look in to software based design
of auto parts. And also the required to make all
communication online. The company does not have an
overhead to create new IT department or purchase high end
computers for designing. What Solutions are available to
SR Autos.
Case Study 1
Case Study 1

Case Study 2

Case Study 3
Solution
Case Study 4

Case Study 5
Migrate to cloud services
Case Study 6 Opt for SaaS services
Case Study 7
Applications will be delivered through web
Case Study 8

Case Study 9
Services are managed by third party.
Further Company will be free from buying and maintaining high
Reading
end systems
Company will be free from upgradation and licensing
cost.
Company designers just need to get authenticated and
access design softwares or other IT services.
Case Study 2
Case Study 1

Case Study 2

Case Study 3
Problem Statement
Case Study 4 A company Balto Co. has many legacy applications. The
Case Study 5 company is running them on dedicated hardware. Company
Case Study 6 is losing money in maintenance of hardware. Balto Co.
Case Study 7
don’t want to change their applications and their need is
Case Study 8
languages and different OS support on the platform.
Case Study 9

Further
Reading
Many companies like Balto Co faces a situation where they
have large costs to run existing applications that they would
like to reduce. Frequently these applications are being
phased out one by one over time but this could take years to
transition and in the meantime the cost of operating all
legacy applications would have to be accepted by the
company. The solution in this scenario comes in the form of
PaaS.
The need of this company can be fulfilled by renting
Case Study 1
resources from a PaaS cloud. However, since their basic
Case Study 2 requirement is to have different applications understandably
Case Study 3 requiring different languages/platforms, it calls for a polyglot
Case Study 4 PaaS. The company needs to choose the vendor carefully,
Case Study 5 ensuring that the specific requirements are met by their
Case Study 6 service that knows how to handle the required types of
Case Study 7 legacy applications they are dealing with.
Case Study 8
However, in spite of this, many of Balto Co’s applications may
Case Study 9 still not run in the PaaS cloud, especially the ones that need
Further
Reading
special hardware. Further codes of many legacy applications
may not run directly on the cloud provider’s hardware and it
is not easy to change the code due to various reasons
ranging from code not being known to the current employees
with the original owner no more with the organization to
having a third party developing the code who may have gone
out of business. Whatever be the case, it is quite possible
that Balto Co may still need to run some of the applications
Case Study 2
Case Study 1

Case Study 2

Case Study 3

Case Study 4
Solution
Case Study 5
Balto Co will go for polyglot (a typical PaasS which
Case Study 6

Case Study 7
supports various languages, and various OS) PaaS
Case Study 8 The Focus is on the operations, the vendor and PaaS
Case Study 9 should be chosen appropriately who has experience
Further
Reading
and capability of handling legacy applications.
The solution offered ,is partial because PaaS may not
offer specific hardware used, codes and environment of
legacy applications may not match exactly used by
company.
Case Study 3
Case Study 1

Case Study 2

Case Study 3 Problem Statement

Case Study 4
P&MCo works on a government application development
Case Study 5
project with another firm SiyaDevelopers, i.e., the
Case Study 6

Case Study 7
developers from both the organizations work on an
Case Study 8
application development project collaboratively. The project
Case Study 9 requires a python platform for development and a versioning
Further system to manage the project. When the application is
Reading
deployed, it should be scalable enough to meet the
demands and must have high availability. But the
developers must concentrate on the problem in hand and
shouldn’t be concerned about the scalability and availability
of the application during the development process. What
must the developers of the two companies do?
Case Study 3
Case Study 1

Case Study 2 They must resort to using a PaaS community cloud, a

Case Study 3 service that must offer a development platform for
Case Study 4 building applications or deployment platform for running
Case Study 5 applications and that can be accessed by the
Case Study 6 developers of both the companies. The middleware
Case Study 7
component in PaaS is responsible for collaborative
Case Study 8
resource management and application scaling based
Case Study 9
on demand of the developers of both P&MCo Company
Further
Reading and SiyaDevelopers
They can work both online as well as offline (for more
sensitive part of the application being developed) and
use the development platform for building applications
or deployment platform for testing the developed
applications. When they are ready, they can deploy the
developed applications in the government’s required
premise
Case Study 3
Case Study 1

Case Study 2

Case Study 3

Case Study 4

Case Study 5

Case Study 6
Solution
Case Study 7

Case Study 8
PaaS community Cloud
Case Study 9 Both partner companies will be able to connect
Further
Reading PaaS middleware will facilitate collaborative and scaling
application, Version control (V 1.0, V2.0, etc.,)
Case Study 4
Case Study 1

Case Study 2

Case Study 3

Case Study 4
Problem Statement
Case Study 5

Case Study 6
A Bank had its own infrastructure setup to manage the
Case Study 7 internal operations like accounts, employee information
Case Study 8 management and other general documents necessary for
Case Study 9 banking. Recently it faced sudden web server crashes due
Further to the inability to meet the spikes in usage of its website.
Reading
The bank cannot completely rely on a public cloud as the
sensitive data cannot be hosted on a third party premise. At
the same time, the private cloud alone cannot
accommodate sudden demand spikes.
Case Study 4
Case Study 1

Case Study 2
Solution
Case Study 3

Case Study 4
Hybrid Cloud.
Case Study 5 Private Cloud for Sensitive Data, Public Cloud for
Case Study 6 non-sensitive Data.
Case Study 7
Interface layer between Private Cloud & Public Cloud in
Case Study 8

Case Study 9
case spike in demand.
Further
Reading The solution lies in having a hybrid cloud. Since bank has
sensitive data, it cannot use a public cloud. Hence it has to
create a private cloud to look after all its requirements.
However, at the peak of demand, they may face a cloud
burst situation wherein they need additional resources.
Under such circumstances, they can plan and use
resources from a public cloud and use these for operations
involving non-sensitive data only.
Case Study 5
Case Study 1

Case Study 2

Case Study 3 Problem Statement

Case Study 4
LC Corp has a data center with a private IaaS cloud set up
Case Study 5

Case Study 6
in Chennai. They manage the cloud well and have a steady
Case Study 7
set of important and happy customers. However, during
Case Study 8 cyclone Vardah, the data center was damaged and it took
Case Study 9 them many days to recover fully. They faced serious
Further customer complaints. Now, during the renewal of contracts,
Reading
one important customer refused to continue with them.
Managers in LC Corp know that they have to do something.
But they already have a proper private cloud and they don’t
want to disturb that. However, their problem is that their
cloud is vulnerable to catastrophic events. What would they
do now to assure their customer?
Case Study 5
Case Study 1

Case Study 2 What they need actually is high availability and disaster
Case Study 3 recovery in their data center. A typical way of doing that is to
Case Study 4 set up an identical data center in a different geographical
Case Study 5 location. This geo-redundant setup would ensure
Case Study 6 uninterrupted service in the face of a similar problem in their
Case Study 7 main data center in future. However, setting up an identical
Case Study 8 cloud elsewhere is an expensive option and not financially
Case Study 9 viable since this needs twice the capital expenditure.
Further LC Corp would continue to keep their production
Reading
environment in their Chennai based private cloud. They
would create a recovery environment in a public cloud and
replicate all data.
They will not use any other resources from the public cloud
under normal circumstances. In the event of a disaster,
administrators can spin up virtual machines in the public
cloud as per the requirements of the customers and can
allow the customers work continue uninterrupted.
Case Study 1
They can relinquish the additional resources from the
Case Study 2
public cloud as and when their own data center is fully
Case Study 3
recovered. This use case shows a hybrid cloud to
Case Study 4

Case Study 5
promote high availability (HA) and disaster recovery
Case Study 6
(DR). This kind of disaster recovery is called a "Warm
Case Study 7 DR" scenario.
Case Study 8
Solution
Case Study 9

Further High Availability and disaster recovery

Reading
A geo redundant set up for their Data Centre
Setting up identical cloud elsewhere would cost twice the
capital cost
Use hybrid cloud for high availability and disaster recovery.
Case Study 6
Case Study 1

Case Study 2
Problem Statement
Case Study 3 Founded in 2010, India-based Chumbak launched its retail
Case Study 4 business selling apparel and home decor items.
Case Study 5 Increasingly, Chumbak sells more of its products via its web
Case Study 6 store. With the web store generating more revenue,
Case Study 7 Chumbak needs to make sure the site continues to meet its
Case Study 8 two goals: delivering a great user experience and
Case Study 9 encouraging customer loyalty.
Further The user experience depends heavily on the infrastructure
Reading
behind the web store. The infrastructure needs to scale
easily so site performance isn’t compromised by sudden
increases in traffic. It’s also important for Chumbak to work
with a cloud provider that offers "pay-as-you-grow" services.
The Chumbak IT team also needs an infrastructure that is
easy to work with. It could not afford for IT staff to be working
on infrastructure maintenance. Furthermore, and also can’t
have developers spending more time on scaling the backend
infrastructure.
Case Study 6
Case Study 1

Case Study 2
Chumbak qualified for AWS Activate, which provides startups
Case Study 3
with low-cost, easy-to-use cloud computing resources
Case Study 4
Chumbak has increased its use of AWS services. This
Case Study 5
includes Amazon Relational Database Service (Amazon
Case Study 6
RDS), which the company uses to oversee time-consuming
Case Study 7
database administration tasks.
Case Study 8

Case Study 9 AWS also now offers services that help Chumbak improve
Further the web store’s user experience. The business uses AWS
Reading
Lambda and Amazon Kinesis to capture and process
web-store clickstreams in real time. When a web-store visitor
clicks on a product, it triggers AWS Lambda, which pushes
the clickstream data into Amazon Kinesis. From there, the
data goes into an Amazon DynamoDB database to find the
visitor’s product history, which then appears on-screen to the
visitor with imagery pulled from from Amazon S3.
Case Study 6
Case Study 1

Case Study 2
Solution
Case Study 3 Chumbak chose AWS because its technology and offerings
Case Study 4 aligned more closely with their IT strategy to utilize most of
Case Study 5
the services provided by Amazon
Case Study 6 Chumbak can now dedicate its IT resources to development
Case Study 7 and avoids routine administration jobs with managed
Case Study 8 services such as Amazon RDS
Case Study 9 About twice a month, Chumbak runs campaigns to drive
Further traffic to the web store, providing visitors with special offers
Reading
on Chumbak products. Using Auto Scaling, the backend
infrastructure to the web store running on Amazon EC2
scales automatically to the level of web-store traffic
AWS "pay-as-you-grow" capability enables Chumbak to
avoid the need to make large IT investments upfront.
Because of the success of the web store on AWS, Chumbak
is considering migrating its business-critical, enterprise
resource planning system to the AWS Cloud
Case Study 7
Case Study 1

Case Study 2

Case Study 3
About the company:
Case Study 4
PayU Group operates in 16 countries across Asia, Central
Case Study 5
and Eastern Europe, Latin America, the Middle East, and
Case Study 6
Africa.
Case Study 7

Case Study 8
The Indian operation - PayU India - is one of the top-three
Case Study 9
payment gateway providers in the country with more than 30
percent market share, comprising more than 300,000
Further
Reading merchants.
PayU India provides more than 70 online payment methods
and aims to match merchants needs with the way
consumers shop and pay.
The business, which focuses heavily on data analytics and
data science, has over 800 employees in India.
Case Study 7
Case Study 1

Case Study 2
Problem Statement
Case Study 3
To thrive in market, the transformed PayU India operation
Case Study 4
needed to deliver reliable, responsive payment services.
Case Study 5

Case Study 6
These services included payment gateways that enable
Case Study 7
merchants to take payments digitally from consumers, and a
Case Study 8
planned product that would enable approved consumers to
Case Study 9
consolidate multiple payments for day-to-day goods and
Further
services purchased online.
Reading
Depending on credit profiles developed by PayU India,
consumers would also be able to defer payments for a short
period and pay later using options such as Immediate
Payment Service (IMPS), National Electronic Funds Transfer,
debit/credit card and net banking, or through an electronic
wallet.
The business needed to consolidate data from both sides to
maximize the use of information to make
Case Study 8
Case Study 1
The PayU India data science team calculated the business
Case Study 2
would need to scale up its existing infrastructure fourfold
Case Study 3
while aggregating all data sources into a single database.
Case Study 4 With new payment consolidation and deferral product-called
Case Study 5 LazyPay-being prepared for launch, the business wanted to
Case Study 6 gain the flexibility to run proofs of concept of various data
Case Study 7 products and services in short timeframe.
Case Study 8 PayU India concluded that it needed to adopt a full-featured
Case Study 9 cloud service to deliver its payment gateway and realize its
Further potential as an online-payments leader.
Reading
The cloud-service provider it selected would need to operate
a data center in India-to comply with legislation requiring
consumer’s data to be retained in India-and operate with
minimum latency of 20 to 30 milliseconds.
The selected provider would need to scale quickly to
accommodate unexpected events, such as the late-2016
spike in consumers using cashless transaction services that
stemmed from a government decision to invalidate certain
high-denomination banknotes.
Case Study 7
Case Study 1

Case Study 2 Solution

Case Study 3 PayU India formally decided to implement a cloud-based
Case Study 4 database and associated systems in the AWS Cloud near
Case Study 5 the end of 2015.
Case Study 6
The data stored and managed in the database would be
Case Study 7
used to launch LazyPay, which would be accessible via a
Case Study 8
separate website.
Case Study 9
PayU India has also moved several production environments
Further
Reading to the AWS Cloud, which the business uses to supports its
merchants and consumers.
PayU India created its architecture in the AWS Asia-Pacific
(Mumbai) Region over a six-week period. During that time,
PayU India engaged leading auditors to review AWS
compliance with the Payment Card Industry Data Security
Standard (PCI-DSS)-a necessary requirement for online
payment providers.
Case Study 7
Case Study 1

Case Study 2
Solution
Case Study 3 AWS has been PCI DSS Certified since 2010. The
Case Study 4 company’s auditors found AWS met all the audit
Case Study 5 requirements.
Case Study 6 The AWS infrastructure operates in an Amazon Virtual
Case Study 7 Private Cloud (Amazon VPC) to provide tiered security,
Case Study 8 since a VPN connects the environment with the colocated
Case Study 9 data center that continues to host dependent systems.
Further PayU India replicates databases within the AWS
Reading
infrastructure, and a subset of the transactions are moved to
an Amazon Redshift data warehouse, where queries are
run to reconcile reports and payments, and understand user
behavior.
The business is also evaluating Amazon Machine Learning
for the key function of determining whether the business
should provide credit to Indian residents who do not have
credit ratings, cards, or bank accounts.
Case Study 8
Case Study 1

Case Study 2

Case Study 3
About the company:
Case Study 4

Case Study 5 Agatsa is a startup based in India that is building affordable

Case Study 6
medical devices and developing an ecosystem of care that
Case Study 7
will help people make informed decisions about their health.
Case Study 8 It currently has 10 employees, including its two cofounders.
Case Study 9 Agatsa has developed Sanket a credit card-sized ECG
Further monitor.
Reading
The device can be held between the thumbs for 15 seconds
to get a clinical-grade ECG. No lead or gel is required.
Sanket has a corresponding app on Android and iOS.
The app displays and records the ECG sent to it via a
Bluetooth-based Sanket device.
Case Study 8
Case Study 1

Case Study 2

Case Study 3 Problem Statement

Case Study 4 Agatsa is an electronic device company and their focus for the
Case Study 5 past few years was on developing Sanket and bringing accuracy
Case Study 6 to their algorithms. Sanket is a custom device that records minute
Case Study 7 vibrations of heartbeats via the patients thumbs, and generates
Case Study 8 an electronic signal that is sent to an Android/iOS app using
Case Study 9 Bluetooth. The mobile app records these electronic signals and
Further sends them to a function in the cloud (currently a Python code
Reading
running in Heroku), which converts this signal to an
understandable pattern of ECG. The data collection and
aggregation strategy was missing in the implementation. They
have a Parse-based implementation on which they store
customer profiles and the PDF files of ECGs. While the raw data
that produces the ECG is not stored anywhere, the PDF files
generated on the phone app are uploaded in Parse.
Case Study 8
Case Study 1

Case Study 2
Problem Statement
Case Study 3 The raw data from which the ECG is deduced is discarded,
Case Study 4 and is not kept anywhere.
Case Study 5 Since raw data is not pushed to the cloud, further analytics
Case Study 6 are not possible on the cloud-especially for historical review
Case Study 7 or predictive analysis.
Case Study 8 Apart from customer profiles, no other information related to
Case Study 9 the devices is available on the server side for processing.
Further
Reading There are no means to find out how many devices have
been active since what time and how many ECGs have been
taken with them.
The information about ECG files is in Parse but no reporting
tools are available for the same.
Parse is being discontinued by Facebook, and Agatsa wants
a more scalable platform on which to store data that can
easily be used for analytics and reporting tools and can be
fed to machine learning systems in the future.
Case Study 9
Case Study 1

Case Study 2
Solution Steps
Case Study 3 Evaluating ECG device capabilities Sanket is a custom
Case Study 4 device that currently has only Bluetooth capabilities and thus
Case Study 5 it can only communicate with a Bluetooth-enabled device
Case Study 6 (Android and iOS mobile phones). An ECG data packet is
Case Study 7 about 90 KB in size with all 8-12 lead worth of data. The
Case Study 8
user needs to hold the device in his/her hand for up to 15
Case Study 9
seconds to get the ECG. The data packet of ECG is sent all
Further
at once to the cloud after the data of all leads is collected
Reading
Determining Hub device connectivity to the cloud The
biggest hurdle was the iOS device connectivity to Azure IoT
Hub as there is no IoT Hub SDK for iOS. A REST API
wrapper (written in NodeJS) was created so that the iOS
device can connect to it for creating device identities and
sending device-to-cloud messages.
Building the end-to-end flow After making both Android
and iOS devices talk to Azure IoT Hub, A simple IoT
Hub-based backend was created as shown the figure
Case Study 9
Case Study 1
Building the end-to-end flow After making both Android
Case Study 2
and iOS devices talk to Azure IoT Hub, A simple IoT
Case Study 3
Hub-based backend was created as shown the figure
Case Study 4

Case Study 5

Case Study 6

Case Study 7

Case Study 8

Case Study 9

Further
Reading

Final Architecture
Case Study 9
Case Study 1

Case Study 2 Solution Steps

Case Study 3
Configuring IoT Hub, API app, and Stream Analytics
Case Study 4
They started writing code for both Android and iOS devices
Case Study 5
and started sending data to IoT Hub in a generic format for
Case Study 6
both devices. Also created a Stream Analytics job, which can
Case Study 7
do calculations over the ECG data (if there is ECG data in it),
Case Study 8 and store it in an Azure SQL Database table. The rest of the
Case Study 9 vitals such as temperature and blood pressure are stored as
Further is.
Reading
Sharing the reports with doctors: Agatsa is using
Notification Hubs to register multiple devices against every
user (doctors or patients) and allow the patients to share the
ECG reports with doctors. The shared message will have
deep-linking links with parameters that will in turn open a
Power BI Embedded report in a web control in a mobile app
(the Power BI Embedded work is in progress). .
Case Study 9
Case Study 1

Case Study 2

Case Study 3

Case Study 4

Case Study 5

Case Study 6

Case Study 7

Case Study 8

Case Study 9

Further
Reading

Sample Screen.
Case Study 9
Case Study 1

Case Study 2

Case Study 3

Case Study 4 Company Profile

Case Study 5
Spotify is an iconic global brand for music and podcasts,
Case Study 6

Case Study 7
delivering customers content they want to hear and artists
Case Study 8
new ways to connect with fans and collaborators.
Case Study 9 It has 271 million Listeners, has 50 million tracks 8+ million
Further per sec requests.
Reading
Spotify brings powerful audio experiences to hundreds of
millions of people every day.
Spotify has continued to innovate its offering, while adhering
to the enduring principles for growing and sustaining a
successful business.
Case Study 9
Case Study 1

Case Study 2

Case Study 3

Case Study 4
Migration Objective
Case Study 5

Case Study 6 Spotify’s leadership and engineering team agreed: The

Case Study 7 company’s massive in-house data centers were difficult to
Case Study 8 provision and maintain, and they didn’t directly serve the
Case Study 9 company’s goal of being the best music service in the world.
Further
Reading
They wanted to free up Spotify engineers to focus on
innovation. They started planning for migration to Google
Cloud Platform (GCP) in 2015, hoping to minimize
disruption to product development, and minimize the cost
and complexity of hybrid operation
Case Study 9
Case Study 1

Case Study 2

Case Study 3

Case Study 4
Migration Strategy and Results
Case Study 5 Spotify invested two years pre-migration in preparing,
Case Study 6 assigning a dedicated Spotify/Google cloud migration team
Case Study 7 to oversee the effort. Ultimately, they split the effort into two
Case Study 8 parts, services and data. For services migration,
Case Study 9
engineering teams moved services to the cloud in focused
Further
Reading two-week sprints, pausing on product development. For
data migration, teams were allowed to choose between
"forklifting" or rewriting options to best fit their needs.
Ultimately, Spotify’s on-premise to cloud migration
succeeded in increasing scalability while freeing up
developers to innovate.
Key Takeaways
Case Study 1

Case Study 2
Gaining stakeholder buy-in is crucial. Spotify was careful
Case Study 3
to consult its engineers about the vision. Once they could
Case Study 4
see what their jobs looked like in the future, they were all-in
advocates.
Case Study 5

Case Study 6 Migration preparation shouldn’t be rushed. Spotify’s

Case Study 7
dedicated migration team took the time to investigate various
Case Study 8
cloud strategies and build out the use case demonstrating
Case Study 9
the benefits of cloud computing to the business. They
Further
carefully mapped all dependencies. They also worked with
Reading Google to identify and orchestrate the right cloud strategies
and solutions.
Focus and dedication pay huge dividends. Spotify’s
dedicated migration team kept everything on track and in
focus, making sure everyone involved was aware of past
experience and lessons already learned. In addition, since
engineering teams were fully focused on the migration effort,
they were able to complete it more quickly, reducing the
disruption to product development.
Further Reading
Case Study 1

Case Study 2

Case Study 3

Case Study 4

Case Study 5

Case Study 6

Case Study 7 1 https://aws.amazon.com/solutions/case-studies/chumbak_india/

Case Study 8 2 https://aws.amazon.com/solutions/case-studies/payu$_$india/

Case Study 9 3 https://microsoft.github.io/techcasestudies/iot/2016/11/29/Agatsa.html

4 https://cloud.google.com/customers/spotify
Further
Reading
Case Study 1

Case Study 2

Case Study 3

Case Study 4

Case Study 5

Case Study 6

Case Study 7

Case Study 8 Thank you

Case Study 9

Further
Reading
Course on Cloud Computing

Case Studies II

Dr. Gitanjali Chandwani Manocha

Assistant Professor
Department of Electronics & Communication Engg
Thapar Institute of Engineering and Technology, Patiala
Case Study 1
Case Study 1

Case Study 2

Case Study 3 Problem Statement

Case Study 4
Siyaram Autos (SRA) is a manufacturing company that
Case Study 5

Case Study 2

Case Study 3
Solution
Case Study 4

Case Study 5
Migrate to cloud services
Case Study 6 Opt for SaaS services
Case Study 7
Applications will be delivered through web
Case Study 8

Case Study 2

Case Study 3

Case Study 4
Solution
Case Study 5
Balto Co will go for polyglot (a typical PaasS which
Case Study 6

Case Study 2

Case Study 3 Problem Statement

Case Study 4
P&MCo works on a government application development
Case Study 5
project with another firm SiyaDevelopers, i.e., the
Case Study 6

Case Study 2 They must resort to using a PaaS community cloud, a

Case Study 2

Case Study 3

Case Study 4

Case Study 5

Case Study 6
Solution
Case Study 7

Case Study 2

Case Study 3

Case Study 4
Problem Statement
Case Study 5

Case Study 2
Solution
Case Study 3

Case Study 2

Case Study 3 Problem Statement

Case Study 4
LC Corp has a data center with a private IaaS cloud set up
Case Study 5

Case Study 5
promote high availability (HA) and disaster recovery
Case Study 6
(DR). This kind of disaster recovery is called a "Warm
Case Study 7 DR" scenario.
Case Study 8
Solution
Case Study 9

Further High Availability and disaster recovery

Case Study 2

Case Study 2
Problem Statement
Case Study 3
To thrive in market, the transformed PayU India operation
Case Study 4
needed to deliver reliable, responsive payment services.
Case Study 5

Case Study 2 Solution

Case Study 2

Case Study 3
About the company:
Case Study 4

Case Study 5 Agatsa is a startup based in India that is building affordable

Case Study 2

Case Study 3 Problem Statement

Case Study 5

Case Study 6

Case Study 7

Case Study 8

Case Study 9

Further
Reading

Final Architecture
Case Study 9
Case Study 1

Case Study 2 Solution Steps

Case Study 2

Case Study 3

Case Study 4

Case Study 5

Case Study 6

Case Study 7

Case Study 8

Case Study 9

Further
Reading

Sample Screen.
Case Study 9
Case Study 1

Case Study 2

Case Study 3

Case Study 4 Company Profile

Case Study 5
Spotify is an iconic global brand for music and podcasts,
Case Study 6

Case Study 2

Case Study 3

Case Study 4
Migration Objective
Case Study 5

Case Study 6 Spotify’s leadership and engineering team agreed: The

Case Study 2

Case Study 3

Case Study 6 Migration preparation shouldn’t be rushed. Spotify’s

Case Study 2

Case Study 3

Case Study 4

Case Study 5

Case Study 6

Case Study 7 1 https://aws.amazon.com/solutions/case-studies/chumbak_india/

Case Study 8 2 https://aws.amazon.com/solutions/case-studies/payu$_$india/

Case Study 9 3 https://microsoft.github.io/techcasestudies/iot/2016/11/29/Agatsa.html

4 https://cloud.google.com/customers/spotify
Further
Reading
Case Study 1

Case Study 2

Case Study 3

Case Study 4

Case Study 5

Case Study 6

Case Study 7

Case Study 8 Thank you

Case Study 9

Further
Reading

Introduction To Cloud Computing
No ratings yet
Introduction To Cloud Computing
107 pages
Module#1
No ratings yet
Module#1
79 pages
Cloud Computing
100% (2)
Cloud Computing
61 pages
Cloud Computing
No ratings yet
Cloud Computing
96 pages
Cloud Computing
No ratings yet
Cloud Computing
3 pages
Cloud Computing Details
No ratings yet
Cloud Computing Details
12 pages
Chapter 1 Introduction To Cloud Computing
No ratings yet
Chapter 1 Introduction To Cloud Computing
60 pages
Cloud Computing
No ratings yet
Cloud Computing
44 pages
Our Seminar
No ratings yet
Our Seminar
26 pages
Unit-2 Cloud Computing
No ratings yet
Unit-2 Cloud Computing
23 pages
IV Btech Cloud Technologies Unit 1
No ratings yet
IV Btech Cloud Technologies Unit 1
66 pages
Introduction To Cloud Computing: Dr. Pummy Dhiman Assistant Professor Cuiet
No ratings yet
Introduction To Cloud Computing: Dr. Pummy Dhiman Assistant Professor Cuiet
178 pages
Cloud Computing Intruducation
No ratings yet
Cloud Computing Intruducation
2 pages
CLOUD 1,2,3 Important Question
No ratings yet
CLOUD 1,2,3 Important Question
14 pages
L1 - Introduction To Cloud ComputingV2
No ratings yet
L1 - Introduction To Cloud ComputingV2
27 pages
Evolution of Cloud Computing
No ratings yet
Evolution of Cloud Computing
22 pages
1.basics of Cloud Computing
No ratings yet
1.basics of Cloud Computing
13 pages
In DR Oduction of Cloud Computing
No ratings yet
In DR Oduction of Cloud Computing
8 pages
Cloud Computing Notes
No ratings yet
Cloud Computing Notes
9 pages
01 A Brief Introduction To Cloud Computing
No ratings yet
01 A Brief Introduction To Cloud Computing
25 pages
CC - Module 1
No ratings yet
CC - Module 1
76 pages
Red Hat Cloud Foundations
No ratings yet
Red Hat Cloud Foundations
12 pages
Cloud Computing Basics & Advanced Guide
No ratings yet
Cloud Computing Basics & Advanced Guide
98 pages
Cloud Computing Unit-1
No ratings yet
Cloud Computing Unit-1
38 pages
Cloud Computing: Benefits and Models
No ratings yet
Cloud Computing: Benefits and Models
18 pages
Cloud Computing Lecture Slide 1
No ratings yet
Cloud Computing Lecture Slide 1
32 pages
CC Unit 01
No ratings yet
CC Unit 01
51 pages
Cloud Computing: Unit 1 Date: 28-01-2025
No ratings yet
Cloud Computing: Unit 1 Date: 28-01-2025
112 pages
Introduction To Cloud Computing: Dr. Hamzeloo Pasargad University of Shiraz
No ratings yet
Introduction To Cloud Computing: Dr. Hamzeloo Pasargad University of Shiraz
12 pages
Module-1 Notes
No ratings yet
Module-1 Notes
34 pages
What Is Cloud Computing?: Simple Introduction To Cloud Com Puting
No ratings yet
What Is Cloud Computing?: Simple Introduction To Cloud Com Puting
32 pages
Introduction To Cloud Computing
No ratings yet
Introduction To Cloud Computing
8 pages
CC Module 1
No ratings yet
CC Module 1
32 pages
Cloud Computing
No ratings yet
Cloud Computing
20 pages
Introduction To Cloud Computing
No ratings yet
Introduction To Cloud Computing
35 pages
Cloud Computing Course
No ratings yet
Cloud Computing Course
12 pages
Cloude - Week1, 2, 3
No ratings yet
Cloude - Week1, 2, 3
72 pages
CloudComputing101 1684945156
No ratings yet
CloudComputing101 1684945156
16 pages
Introduction To Cloud Computing
No ratings yet
Introduction To Cloud Computing
4 pages
CC 1
No ratings yet
CC 1
39 pages
Cloudcomputing Merged
No ratings yet
Cloudcomputing Merged
165 pages
Cloud Computing (Unit - 1) Notes
No ratings yet
Cloud Computing (Unit - 1) Notes
24 pages
Clud Merged PDF
No ratings yet
Clud Merged PDF
275 pages
Cloud and Network
No ratings yet
Cloud and Network
36 pages
Irfan Jalal Bhat
No ratings yet
Irfan Jalal Bhat
19 pages
Cloud Computing Unit 1
No ratings yet
Cloud Computing Unit 1
13 pages
Concept of Cloud Computing-Cloud Public-Private-Introduction To Parallel Spacing.
No ratings yet
Concept of Cloud Computing-Cloud Public-Private-Introduction To Parallel Spacing.
33 pages
Cloud Computing Fundamentals
100% (1)
Cloud Computing Fundamentals
38 pages
Cloud Computing Fundamentals
No ratings yet
Cloud Computing Fundamentals
9 pages
1694438461-Module 1 Cloud Computing
No ratings yet
1694438461-Module 1 Cloud Computing
90 pages
Cloud Computing Unit 1
No ratings yet
Cloud Computing Unit 1
16 pages
CC Unit-2
No ratings yet
CC Unit-2
27 pages
01 A Brief Introduction To Cloud Computing
No ratings yet
01 A Brief Introduction To Cloud Computing
25 pages
Unit 1CloudComputing
No ratings yet
Unit 1CloudComputing
33 pages
MRU EnOfLife
No ratings yet
MRU EnOfLife
4 pages
Exam Tests: Latest Exam Questions & Answers Help You To Pass IT Exam Test Easily
No ratings yet
Exam Tests: Latest Exam Questions & Answers Help You To Pass IT Exam Test Easily
5 pages
Release Notes For Tivoli Workload Scheduler, Version 8.5
No ratings yet
Release Notes For Tivoli Workload Scheduler, Version 8.5
26 pages
Sahil Resume
No ratings yet
Sahil Resume
1 page
HM - DCC and HM7000 - Text-Based Tutorials
No ratings yet
HM - DCC and HM7000 - Text-Based Tutorials
23 pages
SOC Manager
No ratings yet
SOC Manager
7 pages
MTN 3.5G/GPRS Manual Setup Guide
No ratings yet
MTN 3.5G/GPRS Manual Setup Guide
3 pages
Nuclei
No ratings yet
Nuclei
7 pages
Steam Cloud Sync Log
No ratings yet
Steam Cloud Sync Log
36 pages
Edge Computing
67% (3)
Edge Computing
30 pages
What Are The Requirements For A Catalog To Be Assigned To A Functional Area in Manage Functional Area Catalogs UI
No ratings yet
What Are The Requirements For A Catalog To Be Assigned To A Functional Area in Manage Functional Area Catalogs UI
5 pages
IPC365 Mobile Application Manual
No ratings yet
IPC365 Mobile Application Manual
9 pages
Node Group Synch - Workshop
100% (3)
Node Group Synch - Workshop
38 pages
Dapdap High School Computer System Servicing - 10 Module 5 - Week 5 Perform BIOS Configuration
No ratings yet
Dapdap High School Computer System Servicing - 10 Module 5 - Week 5 Perform BIOS Configuration
3 pages
Self Join - SQLZOO
No ratings yet
Self Join - SQLZOO
6 pages
Intrusion Detection System
No ratings yet
Intrusion Detection System
22 pages
Edwin Lai: CELL (650) 521 4141 - E MAIL Portfolio Linkedin
No ratings yet
Edwin Lai: CELL (650) 521 4141 - E MAIL Portfolio Linkedin
1 page
Unit 3 Nis 22620
No ratings yet
Unit 3 Nis 22620
6 pages
5th Chapter 5 - C L Programming
No ratings yet
5th Chapter 5 - C L Programming
65 pages
Attacking Active Directory With Linux
100% (1)
Attacking Active Directory With Linux
20 pages
Test Case For Calculator
No ratings yet
Test Case For Calculator
3 pages
Question Bank - MPMC MID1 - 2021
No ratings yet
Question Bank - MPMC MID1 - 2021
4 pages
Java U4 One Shot Notes
No ratings yet
Java U4 One Shot Notes
65 pages
CS 116 Spring 2020 Lab #05: Due: Wednesday, February 26 Points: 20
No ratings yet
CS 116 Spring 2020 Lab #05: Due: Wednesday, February 26 Points: 20
6 pages
B1722 16 FC6AUsersEN
No ratings yet
B1722 16 FC6AUsersEN
662 pages
H T M L 5: Ankit Singh 1
No ratings yet
H T M L 5: Ankit Singh 1
193 pages
Chapter 3
No ratings yet
Chapter 3
54 pages
Report C & CPP
No ratings yet
Report C & CPP
20 pages
Service Manual: Viewsonic Va1903Wb/Va1903Wm Model No Vs11618 19" Color TFT LCD Display
No ratings yet
Service Manual: Viewsonic Va1903Wb/Va1903Wm Model No Vs11618 19" Color TFT LCD Display
86 pages
3D Face Papers
No ratings yet
3D Face Papers
12 pages