Cloud Computing

Unit IV
Chapter 1: Infrastructure as a Service (IaaS)
Introduction to IaaS:
IaaS is also known as Hardware as a Service (HaaS). It is one of the layers of the
cloud computing platform. It allows customers to outsource their IT
infrastructures, such as servers, networking, processing, storage, virtual
machines, and other resources. Customers access these resources on the Internet
using a pay-as-per-use model.

The IaaS cloud computing platform layer eliminates the need for every
organization to maintain its IT infrastructure.

Evolution of IaaS
IaaS is the bottom layer of cloud computing service model stack. This layer is
situated right below the PaaS layer and two steps down from the top most layer
of SaaS. IaaS is considered as a collection of physical assets including processors,
servers, operating systems, networks, security, storage, computing and virtual
resources. It helps in setting up a virtual server in very less time and you need to
pay only for the used resources.
It increases accessibility of infrastructure by using technology comprising
servers, storage among other peripheral devices. It can be used with services for
operating systems too. The IaaS model is focussed on the enhancement of

1
technology. It provides direct services and infrastructure components like storage,
network, and so on. An example of IaaS is Amazon EC2.

Mainframes (1950)

Virtual Machines (1970)

Telecommunication
companies (1990)

Cloud Computing

Amazon AWS (2002)

Amazon EC2 (2006)

Mainframes (1950)
The evolution started gradually way back in the 1950s with the advent of
mainframe computing. The plan of using mainframe was to promote resource
sharing and optimum utilization of the devices. This computing helped multiple
users in accessing a central computer (mainframe) through dumb terminals. The
foremost idea behind resource sharing is to reduce the overall infrastructure cost.
The cost to purchase and maintain one mainframe computers for every employee
was not practical for any of the organizations.
Virtual Machines (1970)
Around 1970 the concept of VMs was developed which took the popularity of
shared access mainframes to next level. Virtualization software for example
VMware permitted distinct computing environments to reside on single physical
environment. One or more operating systems could be simultaneously executed
in an isolated environment.

2
Tele Communication Companies (1990)
In 1990s telecommunications companies started offering the users a shared access
to the same physical infrastructure through virtualized private network
connections. Earlier the telecommunications companies used to offer dedicated
point-to-point data connections to individual users. The virtualized private
network connections on another hand provided same service quality as their
committed services at a reduced cost.
Emergence of AWS in the year 2002 is considered to be an important milestone
for the expansion of cloud market. Through its Amazon Mechanical Turk, the
company provides a suite of cloud-based infrastructure services which not only
provides simple computation and storage facilities but also advanced features of
human intelligence. In continuation to this a commercial web-based service was
launched by Amazon by the name Elastic Compute Cloud, popularly called as
EC2, in the year 2006. The emergence helped individuals and small companies
to rent computers at affordable prices and to run their own applications on them.
IaaS Architecture
When we talk about cloud economy, there are four main parts to it which are
infrastructure, applications, platforms, and business processes. Among them
infrastructure is the part which is the starting point for business which search for
new revenue opportunities.
The fundamental component of infrastructure is workload. Workload is defined
as the amount of work load that a single server or an application container can
provision with the amount of allocated resources to it. The allocated resources
consist of processing which are CPU and RAM, data and disk latency along with
networking.

Workload representation

3
There are some limitations to when it comes to cloud architecture and that is why
the amount of resources which can be added, are limited. Large clouds solve this
particular problem by a process called sharding or podding. This is explained as
partitioning the data called pods in a horizontal manner in a database or searching
engine for that matter.

Usually an application can be fit in a single pod only at a time. To maximise

output, the pods should be maximum and aggregation should be of the optimum
level where you are able to collect small pods into larger pools. In the case of
infrastructure cloud, such kind of aggregation is done at the level of Cloud control
system (CCS).
Storage architecture in the pod infrastructure plays a very important role. The two
important storage architectures discussed are Direct Attached Storage (DAS)
and Storage Area Networks (SAN).
SAN Model
The Direct attached storage (DAS), each and every pod is quite large because
every node in that pod has its separate storage. Some challenges of this model is
that since every node has its separate storage facility, that makes the provision of
common storage system very difficult hence features like live migration is
difficult.
Coming after the DAS model, the SAN model supports centralization. It makes
live migration possible on one hand while on the other, increases the capacity for
backups and similar technologies. It however comes with its own set of
limitations which are, the size of the pods need to be small because the SAN also

4
has some limit on the scale of cloud nodes and if we go beyond that limit, then
the architectural bond can break.

After considering the storage architectures, the cloud control system should have
certain qualities like good integration ability with external system, good
monitoring, and capacity management to result in optimised resources.
IaaS Providers
Many high cost, off-the-shelf capable commercial vendors like Amazon, Google,
Microsoft and low-cost open source offerings like Eucalyptus, CloudStack are
available. The categories differ in the nature of services they deliver to the user.
Appropriate selection of IaaS cloud provider depends on several factors such as
business requirements, budget, and availability of skilled resources, scope of
business cloud strategy and the present virtualization environment. Out of many
service providers few popular ones are discussed in the following section.
1. Amazon AWS
AWS is a subsidiary of [Link] which was launched in the year 2006.
It aims at providing on-demand; usage-based billing cloud computing
platforms to companies, individuals, and governments.
Amazon AWS offers a wide range of more than 90 services that includes
compute, storage, database and application services, deployment,
networking, management, mobile, analytics, tools for developer’s and
Internet of Things (IoT). These services are accessed over the Internet
5
 using SOAP protocol and are provided through data center’s established at
various places across the globe.
A categorization of services offered by Amazon AWS are given below:
1. Storage Services
a. Amazon Simple Storage Service:
b. Amazon Glacier:
c. Amazon Elastic Block Storage (EBS)
2. AWS Compute Service:
3. Amazon Database services
a. Amazon relational database services.
b. Amazon DynamoDB
c. Amazon Elastic Cache
d. Amazon RedShift
4. Amazon Application Services
2. RackSpace
RackSpace provides multiple IaaS and other managed service solutions
catering to the varied business requirements, both on its own infrastructure
as well as on OpenStack, Microsoft Azure, Google Cloud Platform (GCP),
VMware, and AWS platforms. It offers well supported technologically
neutral, scalable, economical, reliable, hybrid, and multicloud solution
services and is one of the only providers to support Azure stack. The design
of RackSpace supplies the customers with four types of clouds: public,
private, dedicated servers, and Hybrid.
3. Cloud Stack
CloudStack is an open source, flexible software to provide cloud
orchestration platform originally developed by [Link] in the year 2010.
In 2011 [Link] was purchased by Citrix which later donated the project
to Apache Software Foundation in 2012.
Apache CloudStack is a top-level turkey solution that offers reliable and
scalable multi-tenanted platform for private, public, and hybrid IaaS cloud
by teaming various computing resources. CloudStack provides
comprehensive set of features most organizations expect with an IaaS
cloudlike powerful APIs, rich, and secure management user interface,
account separation, multiple role support, on demand virtual data center
hosting, and so on.

6
 Advantages and Disadvantages of IaaS
1. Reduced Cost:
IaaS provides on-demand computing resources at a fixed cost, which can
be used to run applications and get certain services done. It is beneficial for
companies that have high demand for computing power but do not have
the budget for buying physical infrastructure. IaaS also reduces the time
and money spent on managing the infrastructure, as well as IaaS can also
be utilized to host websites, web applications, and online stores.
2. Improved Security:
Cloud computing offers many benefits for business, including improved
productivity and reduced costs. However, one of the most compelling
reasons to move to the cloud is the added security it offers. When your
business data is stored in the cloud, you have the ability to access it from
anywhere, on any device. This provides you with the ability to work from
any location, at any time, with no local reliance on IT.

3. Disaster Recovery (DR):

Disaster Recovery (DR) is the process of recovering from a natural or
human-caused event that affects a local area, organization, or data center,
and the ability of the organization to recover data and services within a
given amount of time. DR is a critical function of all IT environments, but
especially of cloud environments, where service provisioning is performed
over a network.
4. Uptime:
The importance of uptime cannot be overstated. It refers to a cloud
service‘s capacity to be available to your company 24 hours a day, seven
days a week. Cloud companies can guarantee you more than 90% uptime,
and the only time they can’t is if their own server, network, or data centers
go down. This is advantageous to any company.

7
 Disadvantages of IaaS:
1. Security Risks:
It is easy for hackers to break into a cloud computing environment because
it is not as secure as a physical server. If the system is on a local network,
then it is far more secure. However, this advantage of a cloud
computing environment also has its disadvantages, such as lack of control
over the resources.
2. Unexpected Cost:
IaaS is typically cheaper than buying the same services in a packaged
format. However, the cost of using IaaS services can increase
unexpectedly, which can make IaaS appear even more expensive than
traditional services. For example, IaaS providers typically use a formula to
calculate how much bandwidth you use each month. If your usage
increases, your bill can increase even if you’ve never used more services.
3. Dependency on Third-Party Service Provider:
IaaS is a great choice for many organizations, but it comes with its own set
of challenges. One of the biggest challenges is how dependent you are on
the provider. In the case of Amazon Web Services (AWS), the company
owns the hardware, so it is the sole provider of your computing resources.
When you use IaaS, you are paying for the computing resources in addition
to paying for the operating system, software applications, and data storage.
4. Technical Problems:
The technical problems of IaaS are caused by the lack of a standardization
of the cloud environment. Cloud service providers have not yet
standardized on the following aspects of their cloud environments:
networking, service models, security, and network architectures.

8
 Chapter 2: Data in Cloud

Cloud (computing) refers to a set of capabilities that provide infrastructure

(servers and storage) as well as services (e.g. machine learning, deployments,
etc.) that can be provisioned over the internet, usually via APIs.

The data is stored and accessed in cloud via Internet which reduces the hardware
and software demands on the user’s side. This phenomenon of data storage in
cloud is known as cloud storage which hands over your data to a provider (storage
service provider) who stores it in their Data centres.

Evolution of Network Storage in cloud

Increase in data and the need to keep it secure, persistent and consistent
necessitates organizations to improve methods and techniques to manage and use
their data. The limitations in the conventional storage model became more
benefits to cloud computing. To meet the demands that cloud computing imposes
on data storage, cloud technologists turn to a new storage model.

Cloud storage offers a method of storing our resources such that these resources
are easily acquired & accessible, easy to use, and also inexpensive. Cloud, instead
of being a tangible product or a technology is a technique to bring IT resources
in a manner that provides the following key features:
 Self-service
 On-demand availability
 Pay-per-use consumption
The following table depicts the three generations of data storage models, namely,
the Direct Attached Storage (DAS), Storage Area Network (SAN) and the
Network Attached Storage (NAS) models.

9
 Characteristics DAS SAN NAS

Abbrevation DAS stands for SAN stands for NAS stands for
Direct Access Storage Area network attached
Storage. Network. storage.

Description The simplest Multiple computers NAS includes its own

storage model used use a single pool of physical hardware
by most laptops, storage, called the peripheral referred to
phones & desktop shared storage. as a head. The head
computers. runs a simple OS and
has the job of
connecting to the
local network,
authenticating
requests from servers
and managing the file
operations on the
storage drives.
Storage Type Sectors Shared files Blocks
Access Mode Client or sever Client or server Servers

Capacity on 109 109 to 1012 >1012

bytes
Complexity Easy Moderate Complex

Structure of SAN, NAS and DAS models:

10
 Data as a Service ( DaaS )

The ability to provide datasets / data feeds / access to structured data on demand
- to any customer, regardless of their own infrastructure and geographical location
- would not be possible without the advancements in internet technology and
cloud computing made over the past 10–15 years.

The Data as a Service (DaaS) is a member of the cloud-based computing as a

service family and aims to alleviate the problems regarding data management by
providing a cloud-based data centre. Traditionally, all companies used to store
and manage their own data.

Customer relationship management (CRM), enterprise resource planning (ERP),

e-commerce and supply chain systems are some of the widespread business
applications that are employing DaaS. Amazon, Microsoft, and Google are some
popular DaaS providers.
Architecture of DaaS is also defined using the Gather-Process-Publish framework
as shown in the following figure.

The framework is defined for the following functionalities:

 Gather: In this, data files from various sources and in various formats are
collected for further processing.
 Process: The data gathered is normalized and reshaped. A specialized
view of the data is also created.

11
  Publish: Maps are used in order to extract data from the RDBMS to the
various formats as per users’ requirements.
Pricing Models for DaaS
The pricing models of various DaaS vendors can be broadly classified in the
following two categories:
1. Volume based model: This has two approaches. The pricing can be
quantity based or call based.
a. Quantity based pricing: As the name suggests, the customers are
charged based on the amount of data they use. One of the simplest
approaches, it also provides the facility of availing unlimited data
commonly called as the fire-hose approach.
b. Pay per call: In this, the customer is charged on the basis of number
calls made to the API.
2. Data-type based model: In this, the customer is charged based on the data
or attributes required by the customer. Some examples of different types of
data needed by the customers include historical and geographical data.

Database as a Service (DBaaS)

Database as a Service (DBaaS) is a part of the SaaS category in which a cloud-
based database is provided to the companies which can be accessed to store all
kinds of business data. Companies are freed from the task of managing expensive
servers, loading, booting and updating them on a daily basis.
The database has a component called database manager used for controlling and
managing the various instances through an API which is accessible to the various
users through a management console like a web application.
Some popular DBaaS include Amazon’s Relational Database Service (RDS),
Microsoft SQL Database and Google’s cloud SQL.
Users of DBaaS
Similar to other cloud services DBaaS has two types of users:
1. The organization providing the cloud service: The organization providing
the database as a service to the customers usually acts as the Database
Administrator (DBA) as well. It provides the customers a list of available
databases supported by the organization to choose from which generally
includes all the popular relational and non-relational databases. The DBaaS is
usually configured to support specific releases of the software. It is also the

12
 responsibility of the organization providing DBaaS to setup and adhere to
standard policies regarding timely backups and security to ensure data is
backed up from time to time and only authenticated users have authorized
access to the database.
2. Developers or end users: The end users are provided with an abstracted view
or a portal or interface to choose the database of their choice and all operations
by the users are carried out through the interface only. The exact
implementation and operation details are not shared with the end users.
The DBaaS system provides simple mechanisms for addition of users, creation of
databases and schemas and for granting access rights to various end users as per
the application requirements.
Cloud Based Data Storage
Cloud based storage is a model of data storage in which the propelled data is
secured in reliable pools, the physical storage navigates distinctive servers, and
the physical condition is regularly had and overseen by a hosting association.
These cloud storage suppliers are accountable for keeping the data open and
accessible, and the physical condition guaranteed and running. People and
affiliations buy or lease storage limit from the providers to store customer,
affiliation, or application data. The company which provides services of cloud,
keeps the files in an external server to make them available to the user online.
Architecture of Cloud based data storage
Formally, cloud storage is data storage where digital data has been stored in
logical resource pools and the physical storage happens across multiple servers.
There are three key cloud storage architectures namely, public cloud, private
cloud, and hybrid cloud.
Public cloud storage offers multi-tenant environment which is best for
unstructured data. Here, data is kept in global data center’s and the storage data
is widely spread across different regions. The clients will follow a pay-as-you-
use model like a utility payment model.
The private cloud refers to an on-site storage service that gives a dedicated
environment protected by firewalls; the private cloud is best suited for users
demanding customization and greater data control.
The hybrid cloud is essentially a mix of private and public clouds. The following
figure illustrates the architecture of cloud-based data storage.

13
Storage architecture of cloud consists of the following:
 Front end: It exports an application programming interface to access the
storage. In regular storage systems, this API is nothing but a SCSI protocol;
whereas these protocols are evolving, in the cloud. You can have Web service
front ends, file-based front ends, and even more traditional front end.
 Middleware (Storage logic): Middleware exists just behind the front-end layer
and is termed as storage logic. Features like replication, data reduction are
implemented by this layer.
 Back end: After Middleware, the back end implements the physical storage for
data. This can be an internal protocol that implements specific properties or a
traditional back end to the physical disks.
Advantages and Limitations of Cloud Based Storage Solution
Advantages of Cloud Storage
Cloud based storage obviously have many advantages which are as follows:
 Cost: Whether it is a company or an individual using the cloud-based
services, it has to invest less on operating costs as compared to the ones
who use in-house hosting solutions or external storage devices such as hard
drives. Moreover, recent online polls indicate that the average cost of one
gigabyte of storage space on cloud storage is around 3 cents. Thus, there is
no need to invest any capital on storage devices.

14
  Accessibility: We can access our files and folders in the cloud anytime
from anywhere in the world, provided we have the required credentials and
access to the internet.
 Recovery: It is one of the chief advantages of using cloud storage is that
an individual at all times has a backup solution if something goes wrong.
 Syncing: If the subscriber has committed any changes to a file or many
files, the changes will automatically be reflected and synced across all
affiliated devices by the cloud itself.
Disadvantages of cloud-based storage
Bandwidth: Numerous cloud storage service providers have bandwidth
allowance to a specified limit. However, some providers of cloud allow unlimited
bandwidth. This factor should be considered by a subscriber when looking for a
cloud storage provider.
Accessibility: You have access to your data as long as you have an internet
connection. As cloud-based storage solutions depend highly upon your Internet
speed, having a low latency can hamper you from accessing the data in real time.
Data Security: There are concerns with the privacy as well as the safety of the
data stored remotely. The possibility of private data of an organization merging
with other organizations data makes some businesses uneasy.
Privacy Issues: After moving the data to a third-party cloud storage provider, a
possibility of an issue concerning who owns the information arises. Is it the cloud
storage provider, or the company, as a client.
Cloud Based Data Storage Service Providers
The Cloud Storage Providers (CSPs) are accountable for keeping the physical
environment protected and running and the data available and accessible. People
buy or lease storage capacity from the cloud service providers to store user,
organization, or application data. It has been surveyed that about 21% of
industries use cloud for computing, while 25% are using cloud storage services.
This means that there is a very good-sized market for the cloud, and especially
for cloud storage.
 Amazon Web Services: Amazon Web Services is considered a leader of
cloud storage in industry, similar to other aspects of cloud computing. The
services it uses are Simple Storage Service (S3) and Elastic Block which are
for basic object storage and for large storage volumes respectively. AWS is an
innovative company which is continuously releasing products and services. It
has a wide dimension of cloud storage features and services.

15
 Google Cloud Storage: Google Cloud Storage is the underlying storage
service for the company’s other cloud services and products, Data in Cloud
including Google App Engine, Google Compute Engine, and BigQuery,
which are nothing but cloud-based virtual machines and a big data analysis
tool, respectively.
 HP: HP made announcement of its cloud storage platform debuted in May of
2012 and HP offers free 24/7 call and chat support on top of 99.95%
availability. HP is well-positioned to understand enterprise IT storage
requirements, because of its considerable hardware, software, and service
offerings. The system automatically duplicates data across three available
zones for providing resiliency.
 Microsoft: Microsoft’s windows Azure Blob Storage is the second most
widely-used cloud storage service. Microsoft is called a fast follower of
feature roll outs behind Amazon as well. Its support options are very attractive
to large enterprise customers, providing customers a hands-on fee-based
support team. Microsoft recently purchased StorSimple, a cloud storage
gateway vendor to fill its storage chops too.
 Nirvanix: Nirvanix is for companies looking for data-oriented storage needs,
but could be a negative point for customers looking for an all-inclusive vendor
that would provide computing services on top of a storage platform.

16