CLOUD COMPUTING

UNIT-III
Infrastructure As A Service(IAAS) & Platform and Software As A
Service(PAAS/SAAS)
The cloud is an emerging and one of the most popular computing
technologies today. It has capability to serve millions of users at a time over the
network.

o Every technology has some concepts which forms a base over which
technology works/functions.

o Not only this, every technology is also associated with some components
and processes/principles over which technology is built.

o The Cloud Computing technology is also not an exception, it also includes

some components over which cloud technology is built and has some other
components that works over it.

o To understand any technology in deep, first there is a need to understand its

basic principles/components thoroughly.

The most important components of the cloud are

 Cloud Architecture
 Cloud Anatomy
 Networking in cloud.
 Applications on cloud
 Cloud Management
 Cloud Infrastructure Management
 Cloud Application Management
 Migrating Applications to cloud
 Phases of Application Migration
These components need to be understood thoroughly before analyzing/studying
cloud Technology in deep.

Dr. P. Sammulal, Professor of CSE, JNTUH

1. Cloud Architecture:

Every Technology contains an architecture over which technology works.

The architecture provides hierarchical structural over view of the technology and
describe each component and its working.

Cloud computing technology is also contain an architecture that describes about

the components of its architecture and their importance and how they are used to
make the technology to work on it

Based on the user access to the cloud, the cloud architecture is divided into 4-
layers

- Layer-1 (User/client layer)

- Layer-2 ( Network layer)

-Layer-3 (Cloud Management layer)

-Layer-4 (Hardware Resource layer)

(i) User/Client Layer(Layer-1):

 This is the lowest level layer in the architecture

 It includes end user/clients of the cloud who initiate the connection to the
cloud for cloud services.

 The device initiating the connection can be a desktop/pc/ tablet /Smartphone

thin client computer/thin client platform based device or it can be any
electronic device that have capability to connect to the Internet and can
access the web application

 The thin client computer is a generalized computer that have low processing
computing power and depends on other computer for its full functionality.

 The thin client base Computers are becoming popular these days. They are
widely used today to connect to the cloud for getting cloud services.

Dr. P. Sammulal, Professor of CSE, JNTUH

  This layer does not come wider Service Level Agreement (SLA). The SLA
is the terms and conditions document accepted by the both cloud service
provider and cloud service user/client. The SLA contain list of the services
opted by the client their cost, service uptime, down time, the strength of
computational power, availability of the cloud services, etc. all these details
are written and accepted by the both user and service provider.

 Once, SLA is signed by both parties, the cloud service provider offers the
services to the user as per the terms and conditions written in SLA
Document.

 Any violation of provisioning of any service or resource will be prosecuted

legally and heavy penalty will be charged on the service provider.

 The cloud services are offered to the user over the network, the user can
access these services as like the web services. There is a lot of difference
between web applications and cloud applications.

(ii) Network Layer (Layer-2):

 This is the most important layer of the architecture.

 The cloud Technology mainly depends on network layer because it offers

all its services to the client over the network

 The network layer allows the users to connect their network devices like
PCs/laptops/Tablets etc. to the network for gaining access to the cloud for
getting Cloud services

 The network that is used to connect it Public Cloud will be the Internet. The
Public Cloud collects various computing resources from different places and
interconnects them over the Internet, makes them work as one unit to solve
the Computational problem.

 When a user is accessing public cloud, he is not having any information

about the location of the resource being offering the service to him, but he

Dr. P. Sammulal, Professor of CSE, JNTUH

 will be having higher level abstract information like the name of the data
center/ country /state from where he is getting the Service.

 The public cloud can be accessed by a user from any place in the world, at
any time by any device.

 In case of Private cloud, the network over which services offered by the
cloud is a LAN or WAN because private cloud is owned, managed and
operated by one organization. Its Services are offered to the employees of
organization owning the cloud.

 The private cloud is located with in the organization premises and it is local
to the organization network. The private cloud services are not accessed
from outside, they are available with in the organizational network.

 The network layer does not comes under SLA because of the factors
effecting the Quality Of Service(QOS).

 To get good Quality of services either from Public Cloud or private cloud,
the users need to have minimum network bandwidth usually, it is specified
by the cloud Service provider

 The user must use the network with good bandwidth to get good quality of
service, But it is completely left to the choice of the user. Hence it is not
under the purview of SLA.

(iii) Cloud Management Layer(Layer-3):

 The cloud Management layer includes the softwares that is required to

manage the cloud resources.

 The software can be cloud operating system or it is a software that acts as an

interface between cloud user and data center.

 The software provides the facility to the service provider to manage the
cloud resources in terms of its scheduling, optimization, resources
consolidation, storage consolidation, allocating or deallocating the resources.

Dr. P. Sammulal, Professor of CSE, JNTUH

  The complete management of the resources is done by the Software.

 This layer is most important layer in the cloud architecture because it comes
under the perview of SLA.

 Any operations done in this layer will affect the SLA, if the cloud vendor is
not able to provide any Service of resource to the user as per the SLA, it is
treated as SLA violation. if there is a SLA violation, the cloud Service
provider will be charged with penalty. This penalty amount of service
provider must be payed to the user.

 The popular cloud service providers are

o For Public cloud- Amazon Web Services (AWS)

-Microsoft Azure Services

o For Private cloud - Eucalyptus services

-Open Stack Service

 The private cloud Service providers create, deploy, and manage the cloud at
the location specified by the organization based on their requirement.

(iv)Cloud Hardware Resources Layer(Layer-4):

 This layer Includes all the physical resources of the cloud.

 In case of public cloud it can be a data center which is used as a back end.
The data center is a huge collection of hardware resources collected from
different places and connected over the network to work as one unit.

 In case of private cloud, large no. of highly configured systems acts as data
center which is usually located in the organization premises.

 The data centers must give access to the user whenever they try to connect to
the cloud. Hence, it must be available always.

 This layer comes under perview of SLA and the operations taking in this
layer will mostly effects the Service Level Agreement (SLA).
5

Dr. P. Sammulal, Professor of CSE, JNTUH

  If there is any discrepancy in provisioning the required service/application,
the cloud service provider is heavily charged penalty and must pay it to the
user. Hence, most of the service providers use the high bandwidth network
and efficient algorithms to transfer the data at high speed between the cloud
and user.

 Usually, the cloud service provider also specify the minimum bandwidth the
user should maintain for getting good quality of services from cloud.

 All these layers are systemically arranged one below the other
having good separation among them as shown below.

User/Client Layer

Layer-1

Networking Layer

Layer-2

Cloud Management Layer

Layer-3

Hardware Resource Layer

Layer -4

Fig: cloud Architecture

2. Cloud Anatomy:

The anatomy of cloud provides a core structural view of a cloud. Anatomy is part
of cloud architecture but, it is different from it.

The cloud architecture provides a hierarchical view of the cloud whereas anatomy
provides core structural view of a cloud.

Dr. P. Sammulal, Professor of CSE, JNTUH

The architecture talks about the components over which cloud technology is
dependent and the components which work over cloud technology whereas the
anatomy does not discuss about any components, it is completely independent.

The architecture is the overall structure of the cloud whereas anatomy is the part of
architecture.

The anatomy of the cloud is divided into 5-layers, each layer does a specific task.
The layers of anatomy is shown below:
 Applications
 Platform
 Virtualized Infrastructure
 Virtualization
 Physical Resources(Servers/storage etc)

Application
Platform
virtualized Infrastructure
Virtualization
Servers/Storage/Datacenters

Fig: Cloud Anatomy.

(i)Application layer:

 This layer consisting of all applications that are implemented to run on cloud.
Every application can’t be deployed directly on to the cloud.

 To deploy an application on to cloud, the application must acquire properties

of the cloud. If it not possess properties of cloud, the properties must be
embedded into the application, then it can be deployed on to cloud.

 The applications implemented for cloud platform possess the properties of

cloud. Hence, they can be deployed directly on to the cloud.

(ii)Platform:

 A platform is a support or base to run the application. The platform provides

all the resources for the applications to run on cloud.
7

Dr. P. Sammulal, Professor of CSE, JNTUH

  The platform provides an environment where applications can run
comfortably on cloud.
(iii) Infrastructure:
 It is the collection of all logical resources that provides a support for
computational capabilities to the user.
 When an application is running on cloud, it uses all the resources from cloud
infrastructure.
 All required resources are provided to the application by the cloud when it is
running within no time. Once, the application is over, all resources allocated
to it are released and used for other applications.

(iv)Virtualization:

 It is the process / technique of creating logical resources over the existing

physical resources.

 The logical resources are separated from physical resources and both are
made available to the application running on the cloud.

(v)Physical Resources:

 It is the collection of all the physical resources available at different places

and connected over the network that provides computational capabilities to
the user.

 All the resources connected to the cloud are available always to the
user/application. The application uses these resources to perform its task.

4. Applications of cloud:-

*The power of computer can be realized by running the applications on the

computer. If the computer has more computational power, the application runs
fastly. If the computational has little computational power, the application runs
slowly.

*The application running on the computer can be divided into 3 types:

1. Stand-alone application
8

Dr. P. Sammulal, Professor of CSE, JNTUH

 2. Web-based application

3. Cloud based application

a) Stand-alone application:-

*These are the applications, developed to work on individual systems like PCs,
Desktops, and Workstations etc.

*These application were there in the early days of the computers.They can serve
only one user at a time.

*These applications run on the computer by using the computational resources and
features available on that computer. They don’t use data or computational power
from other computers.

*These applications do not depend on the network for their functionality. These are
popular in the early days of computer.

*But, as the time was going on, there was a need for the users to gain access to the
services of the applications running on one computer from the other computer.
This has led to the development of web-based applications.

b)Web-Based applications:-

*Web based applications are different from standalone applications in many

aspects.

*These applications were developed based on the client-server architecture model.

*Server is usually a high-end configured system on which application is installed

and running.

*Client is usually a lower-end configured system used to gain access to the

applications running on server through a network.

*Thus, these applications depends on network to serve the users.

*These applications are capable of serving large number of users simultaneously.

Dr. P. Sammulal, Professor of CSE, JNTUH

*These applications are popular those days, even today also they are widely in use
and become a part of our day-to-day life.
*These applications have some limitations:
-These are developed only for one specific platform.
-These are not multitenant i.e, they can’t be shared among
multiple users.
-They are not elastic hence can’t handle heavy loads.
-They can’t handle continuously varying heavy loads since
they are not elastic.
*They don’t provide quantitative measure of the resources consumed
by the user for generating billing information. Hence, the applications are
temporarily or permanently given full access to the applications.
*They don’t use pay-as-you-go/use approach to generate

billing information.

*All these issues are addressed by Cloud based applications.

c) Cloud-Based applications:-

*Cloud based applications are different from the other applications in that they
posses completely different properties.

*Cloud based applications can also be accessed like web-based applications from
web browser.

*As per the definition of NIST, the cloud application should posses following
properties which makes it different and unique from other applications.
-Multitenancy
-Elastic
-Heterogeneous platform
-Quantitative measurement
10

Dr. P. Sammulal, Professor of CSE, JNTUH

 -On demand service

Cloud Application

Multitenency Elastic Heterogeneity Measured Service On Demand

Multitenancy:-

*This property is considered as one of the most important properties of the cloud
which allows an application or resource to be shared among multiple users with
full independence.

*The independence is a logical independence, it is obtained by creating an instance

of application separately for each user. The modifications done to the data by one
user will not affect the data of other user.

*The logical independence depends on physical independence. The degree at

which logical independence is achieved is dependence on the physical isolation of
the applications on the resources (database, servers etc) on which they are working.

*As long as the instances of applications are working on different parts of the
resources, the logical independence among the users (multitenancy) can be
provided, if the applications are physically close to each other, then multitenancy
can’t be achieved. Thus, the logical independence always depend on the physical
isolation of the instances of the applications.

Elasticity:-

*This is another most important property of the cloud which allows the cloud to
provide better services to its users.

*It is defined as the degree to which system is able to adopt to changes in the
workload by allocating or releasing the resources to the applications such that at

11

Dr. P. Sammulal, Professor of CSE, JNTUH

any point of time, the number of available resources should match closely to the
current demand of the application.

*The elasticity allows the cloud to serve its users in a better and efficient way. It
can handle hundreds of users at a time.

*This property also helps cloud to support rapid fluctuations of load i.e, the
increase or decrease in the number of users accessing the cloud.

*Heterogeneous platform:-

*The cloud supports heterogeneity i.e, the application developed for any platform
can be easily deployed on to the cloud. Hence, most of the application developers
prefers the cloud as their platform to develop their applications. Deploying
applications on to the cloud can be easily done. Once the application is ported on
to the cloud, it can be accessed by any user, from anywhere, by any type of device.

Quantitative measurement:-

*Cloud measures each service or the resource offered to the user to know how
much extent it is used by the user. The billing information is generated based on
usage of the resources. If resources are used heavily, the bill is also heavy, if the
resources are used a little, the bill is also little.

*The cloud uses pay-as-you-use/go model for generating billing information to the
user i.e, the user is charged only for the amount of consumed resources. This
facility is not available in web-based applications.

*On demand service:-

*The cloud is capable of providing application services to the users on demand or

whenever they want.

*The application service is available to the users round the clock without having
any restrictions on the time, place and type of the device and duration of the
service.

*All cloud application services can be offered over the internet to the users across
the world. All services can be accessed over the internet through a web browsers.
12

Dr. P. Sammulal, Professor of CSE, JNTUH

*All these are the few properties of the cloud applications. They make cloud as the
unique platform to develop and run applications on cloud.

4. Networking in Cloud:

 The cloud computing Technology is a technology that collects all the

computational resources available at different geographical places and
connects them over the Internet and makes them to work as one unit for
solving computational problem.

 The cloud also allows the resources to be shared among multiple clouds.

 When a application is submitted to the cloud, cloud allocates resources to the

application so that application will smoothly runs on cloud and gives output.
Thus, the interconnections of these resources are very important in deciding
the performance of the cloud.

 The cloud has different deployment models, the networking connectivity of

the resources and their accessibility in every deployment model is important.
Because the network connectivity of the resources and their accessibility
will decide the performance of the cloud.

a) Public Cloud Access and Networking:

 A public cloud is a collection of all the resources available at different places
and connected over the internet and working together to solve a
computational problem.
 The public cloud can be accessed over the Internet or Some Vendors also
provide Virtual Private Network (VPN) connection over the internet to
access public cloud.

 The public cloud which is accessed over the internet is always creates issues
in terms of its Security that will intern effects the performance of the cloud.

 The possible approach for security is to use encrypted tunnels to transfer the
information between user and the cloud. But, when encryption techniques
are used to encrypt data, it introduce delays in the transfer of information
from one place to another but the information is Secured when transmitted
over the internet.
13

Dr. P. Sammulal, Professor of CSE, JNTUH

  To reduce the delays without compromising for security, best routing
algorithms can be selected that reduces the no of hops between the ultimate
Source and destination in end-to-end connectivity.

 When the no of hops are reduced between Source and destination, the
distance is reduced and delays can be minimized.

 But, selecting best or shortest path involves, talk/discussions/or cooperation

with no. of Internet service Providers (ISP) who connects their networks to
form Internet. If this is achieved, the delays can be reduced.

(b) Private cloud Access & Networking:

 As the name represents, Private cloud is owned, managed and operated by

only one organization. Usually is located in the premises of the
Organization.

 The access to the private cloud is given over the LAN or WAN of the
organization. Some network operators establishes VPN over organization
network and provides access to the cloud.

 Since, the organizational network is protected from the outside access by the
other users, there will not be any threat to the data transmitted between the
user and the cloud over the organization network. Thus, the connectivity to
the cloud is always internal to the organization.

 As long as organization network is protected from outsiders, the cloud

performance is not going to be effected because of organizational network.

c) Intracloud Networking for Public cloud services:

 As the name represents, Public cloud is a collection computational resources

available at different places and connected over the internet and working
together to solve computational problem.

 The resources of public cloud usually spread over large geographical area
and connected together on Internet.
14

Dr. P. Sammulal, Professor of CSE, JNTUH

  Thus, the connectivity of these resources plays a major role in providing the
services to the user.

 The resources must be made available to the user when they request for it.

 The interconnectivity of resources in public cloud is very important. These

connections are not visible to the user. The Security of information
transmitted among these resources in the cloud is also important.

 Because all these resources are connected on public network called Internet.
The user may send/receive valuable information to store or retrieve from
cloud. If this information is hacked, it creates problems to user. Hence,
security issues inside the cloud is to be taken care.
 The accessibility mechanisms used to inside the cloud is also important. The
accessibility mechanisms must provide the access to the resources whenever
user demands for it. Thus, the internal network Connectivity among the
cloud resources will effect the performance of the cloud.

 The quality of Service (QoS) among the connected resources is another

aspect. Good Quality of service can be obtained by using Secure network
connections and efficient algorithm for access and data transfer.

(d) Intracloud Networking for Private Cloud Services: A private cloud is owned,
managed and operated by one organization. It contains many computational
resources in the form of highly configured systems.

 These systems act as a data center for private cloud.

 A organization may have one or more data centers.

 All data centers are connected using organizational network connectivity.

How much amount of bandwidth is supported by organizational network to
connect all the resources of cloud is an important issue. Because the network
connectivity must support the applications running in this environment to
give the best service to the user.

15

Dr. P. Sammulal, Professor of CSE, JNTUH

  Since the private cloud is a part of the organization network, it uses this intra
cloud network to connect the users with applications running on cloud. Once
the connection is established, the data flows between the user and the
application over organization network. This network must have sufficient
bandwidth to support for best service.

 If the cloud application was designed based on Service oriented architecture,

the application equally distributes its traffic in all the directions of the
network to provide good quality of service.

 Usually, Organization network is its own network and do not have access to
the outsiders and uses best security policies to protect its own traffic on its
network, equally distributes its network traffic in all possible paths, it does
not affect the performance of cloud as long as its network maintenance is
good, otherwise it impacts cloud performance.

 As the time goes on, the cloud based applications are increasing, many users
are accessing these applications, and slowly it is increasing the traffic on the
network.

 Hence, there is a reason to consider network connectivity and their access

mechanisms to the resources with the increased traffic. As the traffic is
increasing, the network requires more flexible architecture to support
increased load on the network with global standards..

(e)New Facets in Private Networks: Most of the private networks are

conventional/traditional networks designed to deliver Services with the
applications running on the servers located within the premises of the organization.

 The most popular network based applications like E-Mail, File sharing and
Enterprise Resource Planning (ERP) are deigned to work on Servers located
within the organizational network to serve the users.

 Now a days, most cloud Service providers offering their software to the
corporate organization as Software As A Service (SaaS). This software runs
on the cloud owned by third party but, providing services to the
organizations.
16

Dr. P. Sammulal, Professor of CSE, JNTUH

  When the organizations are getting cloud based services, the important issue
is whether the organization Conventional network can support traffic from
the cloud services or not, because these networks are not designed for this
purpose.

 Since, many users in the organization access cloud based services

simultaneously, it increases traffic on organization network, most
organizations network are not architected to optimize performance for cloud
applications.

 Hence, it brings lot of challenges in accessing and usage mechanisms of

applications running on cloud data centers.

 Today, most of the widely used applications are slowly moving from on-
premises based applications to cloud based applications and offering their
services to the organizations as SaaS at inexpensive prices.

 Hence, there is a need to change the architecture of conventional networks

owned by organizations to give the support for cloud based applications.
otherwise, the organizations can’t function if access is not given to cloud
applications as and when needed.

(f) Path for Internet Traffic:

 With the increase of cloud based applications, the no.of users accessing
these applications are increasing day by day, which in turn increasing the
network traffic on the Internet.
 But the Internet is using limited set of gateways to control the traffic. This is
not sufficient to control the increasing traffic on the Internet, it leads to
availability, Performance and access issues to the cloud.
 To overcome the problem, there is a need to keep more no. of Internet
Gateways on the internet especially, at the places where more traffic is found
on the internet.This will address the problem of availability, performance
and access to the cloud and reduces network latency.
 Today, the most popular application running on Internet i.e video
conferencing, which consumes more internet bandwidth and most critical
application of organizations ERP, Consumes less bandwidth on internet.

17

Dr. P. Sammulal, Professor of CSE, JNTUH

  Because of video Conferencing, the ERP application should not be
interrupted or stopped. Therefore, a good and correct plan is required to
divert the traffic equally in various directions on Internet. So that no
application is getting effected because of non-availability of network facility.

 These challenges must be consider on Internet to manage the traffic with

more no. of Gateways and with increased network bandwidth otherwise,
cloud services performance is being affected.

4. Managing the cloud:-

 The cloud management is one of the most important job of the cloud. The
cloud performance mainly depends on how a cloud is managed. If the cloud
is managed properly, it provides a good quality of the services(QoS),
otherwise, it effects the performance of the cloud.
 The cloud management consisting of 2 parts.
-Managing the cloud infrastructure

-Managing the cloud applications

a) Managing the cloud infrastructure:-

 Cloud infrastructure is considered as backbone of the cloud. It is responsible

for providing good quality of services(QoS) to the cloud users. The cloud
performance also mainly depends on the quality of cloud resources.
 In a cloud, cloud resources includes computers, servers, storage, networking
devices, databases, applications, etc. all these resources are collected from
different places and made as a part of cloud infrastructure.

18

Dr. P. Sammulal, Professor of CSE, JNTUH

  The main factors to be considered as per as cloud infrastructure
management is concern are
-Cloud OS
- Cloud performance and SLA
-Cloud maintenance cost
-Power consumption
-Service delivery methods and load fluctuations
-Cloud governance
Cloud OS:
 A cloud contains all its resources from different places and all of
them are connected over the internet.
 The cloud maintains all these resources, provides co-ordination among
them through a cloud core software called Cloud Operating
System(OS).
 The cloud OS maintains all cloud resources by performing tasks such
as resource scheduling, allocating and deallocating resources to the
applications, load balancing, optimizing resources consumption, etc.
 The cloud OS also allows the users to share cloud resources as long
as there is no risks involved.
 The cloud OS must be powerful to efficiently manage its resources,
otherwise the cloud cannot provide give/provide best services to users.
Performance of the cloud and SLA:

 The performance of the cloud is one of the most important feature. It mainly
depends on its resource management.
 If the performance of the cloud is not good, it will lead to several
inefficiencies in terms of QoS and SLA.
 The Service Level Agreement (SLA) is a set of rules document
governed by both the cloud service vendor and the user. This
document is created at the time of purchasing the cloud services from
the cloud vendor by the user.

19

Dr. P. Sammulal, Professor of CSE, JNTUH

  If cloud performance is not good, it can’t satisfy SLA. If SLA is not
satisfied the cloud vendor is charged with heavy penalty, it should be
paid to the cloud user.
 As per as SLA, the cloud should always provide certain basic services
to the cloud user otherwise, the whole purpose of the cloud is last.
Cost of cloud Maintenance:
 The cloud resources must be maintained properly otherwise, it effects
the QoS and SLA.
 If the cloud vendor invests more in maintaining cloud resources, it
will automatically increases the cost of cloud services. If the cloud
service is expensive, many users do not opt for cloud services, this
will results in lower users pool. If the no. of users are less, the cloud
vendor does not get profits.
 On the other hand, if the cloud vendor invests less in cloud
maintenance, it will decrease the cost of cloud services, the services
becomes cheap. If the services are cheap, more user will opt for
cloud services, that increases users pool, if users pool is more, vendor
can get profits but, the cloud can’t offer good QoS. Hence, users will
not be satisfied with cloud services. That leads to the collapse of the
cloud within a short span of time.
Hence, there should be always a trade-off between the investment made
by the cloud vendor in maintaining cloud resources and the QoS offered
to the cloud users.

Power consumption:

 The power consumption of cloud resource is another important issue

to be considered. It will also lead to increase the cost of cloud
services.
 If more number of resources are employed to provide a service to the
user, it consumes more power. In order to reduce more no.of
resources, the techniques like “server consolidation” and “storage
workload” can be used.
 The server consolidation is a process of reducing no.of servers or
server locations that are required by an organization to deliver the
service to it.
20

Dr. P. Sammulal, Professor of CSE, JNTUH

  The storage workload is a process of distributing the workload equally
to all storage units such that no unit is left under utilized or over
utilized.
Service delivery methods of Load Fluctuations:

 All the above factors discussed related to the physical resources

management. Apart from this, there are different management methods
are available to deliver services to the user.
 Different type of services use different service delivery management
methods to efficiently deliver the services to the user.
 All service delivery methods are developed based on “Load Fluctuations”.
A Load fluctuation is a phenomenon in which the system workload
will change continuously. The workload is nothing but the rapid
increase or decrease in the no.of users accessing cloud services.
 The load fluctuations is the critical issue to be considered as per as
cloud applications are concerned.
 Load fluctuations are of two types
- predictable Load Fluctuations
- unpredictable Load Fluctuations
 The predictable load fluctuations can be handled easily. The cloud can be
configured to handle such load fluctuations.
 Unpredictable load fluctuations can’t be handled easily. However, the
methods have been developed to handle such situations by the cloud.
The cloud is mainly intended to handle unpredictable load fluctuations.
Cloud governance:

 The cloud governance is a word widely used in corporate world. It is

closely related to cloud maintenance but it is different from it.
 The cloud governance mainly focuses on policies, SLAs, cloud performance,
QoS, etc.
 The cloud governance is the process of creating a value to the
organization by creating a strategic objective which leads to the
growth of the organization and helps in obtaining more grip on the
organization.

21

Dr. P. Sammulal, Professor of CSE, JNTUH

  Several aspects are focused by cloud governance as mentioned
previously, out of which SLA is most important.
 The total cloud is governed by SLA. If SLA is not satisfied, the
cloud vendor has to pay hear by penalty to the cloud user.
(b) Management of cloud applications:

 Most of the corporate organizations today looking at the cloud to

move or create their corporate applications from their traditional
environment to the cloud environment.
 Moving applications from corporate to cloud environment brings/poses
lot of challenges to the cloud application developers, the applications
becomes more composite and complex.
 The cloud resources also need to be enhanced both in terms of IaaS
and SaaS.
 As per as IaaS is considered, the cloud computers, servers, storages,
databases, platforms need to be enhanced.
 As per as SaaS is considered, the application softwares like e-mail,
SMS, etc need to be incorporated in the cloud.
 All these issues are addressed by the cloud application management.
 The cloud application management techniques provides possible
solutions to make applications to run on cloud platform without any
difficulty.
 Cloud application management also helps in implementing the
corporate policies like governance, auditing and the environment
management required while deploying applications on to the cloud.
 The cloud application management also
- helps in collecting information about various events, analyzes it
and identifies critical information that needs action immediately such as
increasing no.of resources or adding new services to the cloud, etc.

- provides tools and processes to easily manage applications.

Migrating application to the cloud:

22

Dr. P. Sammulal, Professor of CSE, JNTUH

  It is the process of moving one or more enterprise applications and
its IT environment from corporate traditional environment to the
cloud environment.
 Moving applications to the cloud environment is not easy, it brings lot of
challenges to the application developers.
 Moving applications to the cloud, provides lot of benefits to the
organizations in terms of investment made on purchasing hardware and
software and their maintenance. All this cost is saved.
 Moving applications to the cloud is done phase by phase. The phases
of applications migration to the cloud are as follows:
- Evaluation

- Strategies of migration

- Prototyping

- Provisioning

-Testing

(a)Phases of migration:

1. Evalution:

 To migrate an application from other environment to the cloud

environment, the components available in cloud must be evaluated. To
know that whether the components available in cloud can support the
application or not. The evaluation of components includes cloud
resources such as servers, computational power, storage, platform,
applications, cloud architecture, software licenses, application
architecture, etc. Other components includes risks, security issues,
compliance, SLAs, operational processes, management strategies, etc.
 All these factors are analyzed to move applications to the cloud.
2. Migration Strategies:

23

Dr. P. Sammulal, Professor of CSE, JNTUH

  Based on the evaluation, a particular migration strategy is selected to
move application on to cloud.
 Two types of migration strategies are available
-Hotplug strategy

-Fusion strategy

 In the hotplug strategy, the applications and its data and its
dependencies are isolated and moved on to the cloud. Then, the entire
application is made to run at all once.
 In the fusion strategy, the application is partially moved. In the sense,
only application itself is moved but its dependencies such as data and
software licences, specialized servers, interconnections with other
applications are preserved/maintained in the old environment.
Prototyping:

 The migration strategy is normally preceded by prototyping.

Prototyping is the process of validating and assuring that the
application runs in the cloud environment.
 To prove this, a small part of the application is installed on the cloud
and run with the test dataset.
Provisioning:

 It is the process of providing all the required resources in the cloud

environment to make the application to run in the cloud.
 Normally, the short falls identified in the evaluation are provided
which includes
-servers to be provided in the environment.
-required platform softwares are installed and configured.
-required application software is installed and configured.
-the environment is also configured to match with the application
requirement.
-databases and files are replicated/copied.
-all internal and external integration points are configured.
-operation and management software is setup.
-web servers, services, batch jobs are configured.
24

Dr. P. Sammulal, Professor of CSE, JNTUH

Testing:

 After application is deployed, a port migration test are usually

performed to assure that application is successfully migrated and
working.
 The tests that are usually performed are
-failure and recovery tests.

-performance and load testing.

-scale up and down test.

All these tests are performed against expected traffic load and resource
utilization levels.

b. Cloud Migration approaches:

 It is the process of moving from traditional corporate environment to

cloud environment.
 There are four approaches available. Any one of these approaches
can be used to migrate to cloud.
-Migrating existing applications.
-Start from scratch.
-New separate company.
-Buy an existing cloud vendor.
Migrating Existing Applications:

 The applications running in the traditional corporate environment can

be rebuild or rearchitected to suit for the needs of cloud environment.
This can be done by the top level engineers to develop/introduce new
functionalities. This also requires that new R&D decision to be taken.
This is usually done in several release to have a full fledged
applications.
Start from the scratch:

25

Dr. P. Sammulal, Professor of CSE, JNTUH

  The applications available in traditional corporate environment is
rebuilt right from the scratch by taking the advantage of facilities and
features available in the cloud environment.
 This requires the software architects to have grip on the facilities and
features available in the cloud to develop efficient applications.
 Instead of modifying the applications as in first step, developing
application from the scratch is easy efficient.
 All this can be done even with a small team of engineers who have
full knowledge in the cloud environment.
Starting a new company:

 Instead of moving applications from the traditional corporate environment to

the cloud environment, a new company can be started with latest
cloud applications and platform.
 The required funding and software licensing can be obtained from the
existing company.
 This will benefit the new company in many ways as follows
-If any remarks are there on existing company it can’t be carried
on to new company.
-With new company, new brand can be introduced into the market.
-New company may have new management, new policies, different
way of working hierarchy.
 The most important is, the new company must behave like a cloud
vendor.
Buying an Existing Cloud Vendor:

 The existing company, if it is having large amount of capital, it can

invest in purchasing an existing reputed cloud vendor company, it
will save lot of time to the organization and gets two advantages.
-The leading competitor will not be there in the cloud vendors to
the organization.
-It can touch the ground by running in the cloud space with no
time.
 The most important here is, the cloud company innovation techniques,
operational approaches, management strategies, hierarchy should not be
26

Dr. P. Sammulal, Professor of CSE, JNTUH

distributed but, it can be enhanced with new ideas and objectives on a
phase by phase number.

27

Dr. P. Sammulal, Professor of CSE, JNTUH