ISSN 2319 – 1953

International Journal of Scientific Research in Computer Science Applications and Management Studies

Evolution of Computing-Mainframe to Cloud:

A Systematic Study
Anurag Jain#1, Nitin Arora*2, Prashant Rawat@3, Bhupesh Kumar Dewangan$4, Rajneesh Kumar+5
#1
Virtualization Department, School of Computer Science, University of Petroleum & Energy Studies, Dehradun, India
*2
Informatics Department, School of Computer Science, University of Petroleum & Energy Studies, Dehradun, India
@3
Systematics Department, School of Computer Science, University of Petroleum & Energy Studies, Dehradun, India
$4
Informatics Department, School of Computer Science, University of Petroleum & Energy Studies, Dehradun, India
+5
Maharishi Markandeshwar Engineering College, Maharishi Markandeshwar University, Mullana, Ambala, India
[email protected],[email protected],[email protected]

[email protected],[email protected]

Abstract— Nowadays computing process has become more different machines. It uses middle ware to divide and assign
important and usable. After water, gas, telephone and electricity, parts of a program between different computers. Infinite
cloud computing has become fifth essential utility in our life. It is number of resources can be added or removed in the network
happening, as nowadays every computer related service provider at any time. Like power grid concept, concept of grid
is giving services via cloud. Still, cloud computing is an emerging
technology as many of its issues are yet to be unfolded. The goal
computing has provided access to very massive computing
of this paper is to discuss the evaluation of computing from resources to resolve research oriented problems [5, 6].
parallel to cloud computing, along with cloud computing types,
essential characteristics, architecture and research issues.

Keywords— Cloud Computing, Grid Computing, Cluster

Computing, Cloud Computing Types, Cloud Computing
Architecture, Virtualization, Cloud Computing Issues

I. CLOUD COMPUTING EVOLUTION

Computing signifies those activities, which contains the use
and design of software and hardware for an endless collection
of activities with its meaning changing from one context to
another [1]. To achieve better performance in computing,
concept of distributed and parallelism are used. In parallel
computing, numerous computations are executed in parallel
using big pool of processing units which are member of a one Fig 1 Cloud computing evolution [7]
system in which single memory module is shared among all
processing units[2]. In distributed computing, many
Cloud computing is the consequence of the evolution of
processing units are used. Every processing unit is itself
mainframe, distributed, cluster and grid computing. Statistics
individual system. Physically, they may be at the different or
of evolution of cloud computing has been shown in figure 1.
same locations but under the control of single organization
In grid computing, we do the grouping of resources to cater a
with every unit have its own memory module [3]. In cluster
specific task while in cloud computing we do the grouping of
computing, set of dissimilar or similar distributed or parallel
resources for general tasks. In grid computing, big tasks are
systems are interconnected through high-speed networks and
processed for small time, while exactly reverse happen in
functioning as a single virtual machine to solve a problem and
cloud computing. We handle little tasks for long time. A
are viewed by users as a collection of multiple interconnected
single person does not own grid. Its maintenance is the
machines physically co-located [4].
responsibility of government or non-commercial organization.
In early 90’s Grid computing comes into existence by
While cloud can be owned by private company for
dynamic aggregation of geographically dispersed,
commercial purpose. Concept of virtualization has created the
heterogeneous or homogeneous clusters connected via an
tremendous difference in the resource allocation methodology
internet connection. These groups of clusters were owned by
multiple countries. Basic feature of grid computing is that it of grid and cloud computing. Cloud computing is the business
model of grid computing [8, 9].
divides big tasks into small parts and process them on

IJSRCSAMS
Volume 7, Issue 4 (July 2018) www.ijsrcsams.com
 ISSN 2319 – 1953
International Journal of Scientific Research in Computer Science Applications and Management Studies

In this paper, authors have discussed the basics of cloud computing are service model, deployment model and essential
computing thoroughly along with challenges in adoption of characteristics. Same has been described below in figure 3.
cloud computing. Cloud computing definition, characteristics,
architecture and types have been explained in section II. In
section III, virtualization and its types have been discussed in
detail. Research issues and challenges in the area of cloud
computing have been discussed in section IV. Conclusion and
future directions are given in section V.

II. INTRODUCTION TO CLOUD COMPUTING

Internet users are putting a huge amount of data over
servers as a result of rapid advancement of technology. More
computing resources are required in order to process this huge
volume of data.

Fig 3 Visual definition of cloud computing [12]

Cloud computing implies facilities provided via the internet.

These facilities include storage services, application services,
and hardware services. The user can use these facilities
without human interface via the internet irrespective of time
and place. A group of resources is formed. These hardware
and software resources can be varied (increased or decreased)
according to requirement through virtualization. These
services can be accessed by the user by forming their account
and they have to reimburse according to their usage [13]. A
grouping of parallel and distributed connected systems which
are dynamically giving services using virtualization to the user
as per the SLA (service level agreements) documented
between the user and CSP (cloud service provider) [14].

B. Essential Characteristics
All types of cloud model must possess some essential
characteristics which are as follows:
1) Scalability: Scalability indicates the efficiency of the
Fig 2 Cloud computing overview [11] system in handling the increasing demand. Diversified users
from diversified locations having dissimilar time zone may
Simultaneously, the companies which have already made
use the same kind of services at the same time which leads to
huge investments for making a pool of resources are now
overloading of the system. This increasing demand can be
taking a step back by a feeling of less return and lesser
efficiently handled by scaled up cloud resources.
utilization of resources. These factors and emergence of
virtual technology has given a birth to a new model of 2) Inexpensive: Huge investment is not needed while using
computing called cloud computing [10]. Figure 2 shows an the services through the cloud. The user can enjoy the services
abstract view of cloud computing along with its three as per the service level agreement by creating an account with
deployment models. cloud service provider, and they have to pay in accordance
with their usage only.
A. Definition
3) Elasticity: The capability of the system to increase or
Cloud computing is a kind of model, that allows the on- decrease resources as per the dynamically changing need of
demand access to joint group of resources (hardware, users is called the elasticity.
application, storage) at any time and place without human
interaction [12]. Key aspects used by NIST to describe cloud 4) Virtualization: The key concept behind the success of
cloud computing is Virtualization. Resources can be divided

IJSRCSAMS
Volume 7, Issue 4 (July 2018) www.ijsrcsams.com
 ISSN 2319 – 1953
International Journal of Scientific Research in Computer Science Applications and Management Studies

or combined and multiple queries can be handled in parallel

through multiple virtual interfaces [15].
5) Large scale infrastructure: There are approximately one
million servers in Google's cloud platform. Also there are
more than one thousand servers in Yahoo, Amazon, IBM,
Microsoft and other cloud services platform. Even in some
private cloud, there are thousands of servers. This implies
high computing capabilities are provided by Cloud computing
environment to the client.
6) Universality: The facilities provided by the CSP (cloud
service provider) are independent of the application. The user
can select a range of applications as per his needs. Same
environment can be used to run a range of applications by
different clients. Also, the user can access the application just Fig 4 Cloud service delivery model [17]
by internet connectivity irrespective of location and time.
7) Large scale infrastructure: Big software giants possess a Figure 4 shows the above mentioned cloud service delivery
huge pool of resources. Number of servers has further model in the form of layer. These three vital categorizations
increased through virtualization. The huge resource pool of are generally called “SPI” i.e. software, platform, and
Amazon, IBM, Microsoft, and Yahoo has turned large-scale infrastructure respectively.
computing facilities available on the cloud into reality [16].
D. Cloud Deployment Model
C. Cloud Computing Architecture
Based on the way cloud facilities are given to the users,
Computing capabilities provided by cloud service provider cloud deployment models can be classified in four types:
are the services given by cloud. It consist of environment,
1) Public Cloud: Public cloud implies anyone can access the
application, and storage to create the applications and utilize
services. A person willing to use the facilities has to first make
the resources. As an alternative of providing facilities from
his account with the CSP (cloud service provider) and then
one computer, facilities are provided from different servers,
user can access the facilities purchased by him on the pay per
which are connected via internet. Details of the different cloud
usage basis. Cloud service provider (CSP) is liable for
service model are as follows:
providing, maintaining and managing resources.
1) Cloud Software as Service (SaaS): In this service delivery
2) Private Cloud: A private cloud is like public cloud in
model, the user uses the variety of application software
terms of services but different in context of access. A private
through internet. Cloud service provider is responsible for
cloud is planned for a specific group and only the staff of that
installation, license and manage the application software. The
group can use the cloud facilities for group work. Non
user creates their account and pay for the used service. Google
organizational employees cannot use cloud services. The
Apps, Intuit.com, SalesForce.com and SQL Azure are some of
group itself manages all the services of the cloud. There is no
the popular SaaS providers.
issue of bandwidth and security in private cloud.
2) Cloud Platform as Service (PaaS): In PaaS, service
3) Hybrid Cloud: The combination of public and private
provider provides computing environment where the user can
cloud creates hybrid cloud. When the service provider opens a
install and execute their application. Service provider is
portion of the cloud for the public and hide the residual part
responsible for installation, license and management of the
for the staffs of the group then this kind of cloud is called a
platform. The user creates their account and run their
hybrid cloud. Noncritical resources are made open while
application in the environment. IBM Bluemix, Force.com,
critical services are not outsourced.
GoGrid cloud center, GoogleApp engine and windows azure
are some of the popular PaaS providers. 4) Community Cloud: Community cloud includes dissimilar
types of cloud whose services are available for a specific
3) Cloud Infrastructure as Service (IaaS): In IaaS, service
community of users. Users of that community can use the
provider provides storage and computing services to the user.
services available on the cloud [18].
Cloud service provider is responsible for purchase, deploy and
management of hardware resources. The user creates their III. VIRTUALIZATION
account and pays for the hired service. Terremark, Eucalyptus,
Linode, Amazon Elastic Compute Cloud, Flexiscale are some All The discussion of cloud computing basics will be
of the popular IaaS provider [17]. incomplete without a discussion of virtualization.
Virtualization is the key concept behind success and

IJSRCSAMS
Volume 7, Issue 4 (July 2018) www.ijsrcsams.com
 ISSN 2319 – 1953
International Journal of Scientific Research in Computer Science Applications and Management Studies

realization of cloud computing. In this section, authors have Figure 6 displays a contrast between non-virtualized and
discussed the basics of virtualization. virtualized system.
Virtualization is used to hide computer resources. It
includes transformation of physical resources to logical
resource. Resources are divided and shared among multiple
virtual interfaces and all resources are treated as logical rather
than physical. This gives an illusion to the user that multiple
systems are available. Different interface along with resources
are called as virtual machine. System on which virtual
machine is implemented is called host machine.
Diagrammatical viewpoint of virtualization concept has
been shown in figure 5:

Fig 6 Non-virtualized vs virtualized system view

C. Types of Virtualization & Hypervisor

Full virtualization is the type in which the host operating
system and guest operating system have the same capabilities
whereas in para-virtualization the guest operating system has
Fig 5 Concept of virtualization [19] fewer capabilities than host operating system. Hypervisors can
A. Virtualization Advantage be categorized into two categories:

1) Reduce Hardware Cost: Sharing of resources among 1) Type 1: No Operating system is required to run Type 1
different user has helped the investor to get back invested cost hypervisor. Oracle VM, LynxSecure, VirtualLogic VLX are
quickly. Also, this has helped service provider to provide some of the frequently used Type 1 hypervisors. AS type 1
services in a cheap rate. hypervisor is installed on blank system, so it does not have
any operating system. It is better than type 2 hypervisor in
2) Scalable: By handling queries in parallel manner through terms of availability, efficiency and security. The idea of type
multiple interfaces, cloud service provider can handle 1 Hypervisor has been displayed in figure 7.
changing demand efficiently.
3) Flexibility: It allows service provider to give a combined
and relaxed access to resources. [19].
B. Hypervisor
Concept of virtualization is motivated from
multiprogramming concept in which the process thinks it has
full control on all resources. But resources are shared among
multiple processes at run time. On the same line, in
virtualization, the processors are shared between multiple
operating system. Hypervisor/virtual machine manager creates
the illusion that operating system has complete control of the
underlying hardware. A machine which creates, implements
and destroys the virtual machine is called host machine. The
Hypervisor is a low-level program that acts as a VMM
(Virtual Machine Manager). It permits many operating system
Fig 7 Type I Hypervisor [20]
(VM/guest OS) to run concurrently on a lone physical host. It
multiplexes different resources between VM. Hypervisor are 2)Type 2: This hypervisor uses the host operating system
designed for specific processor architecture. KVM, Xen, for its execution. Virtual Server 2005, KVM and Microsoft
VMware, VirtualBox are some of the general hypervisors.

IJSRCSAMS
Volume 7, Issue 4 (July 2018) www.ijsrcsams.com
 ISSN 2319 – 1953
International Journal of Scientific Research in Computer Science Applications and Management Studies

Hyper-V are examples of Type 2 hypervisor [20]. Figure 8 types of resources deployed by the cloud service provider not
displays the idea of type 2 Hypervisor. for the consumed resources. The user is not aware of this fact
as it is changing continuously due to changing demand.
4) Virtual Machine Migration: A company who has made a
big investment in installing and configuring the cloud wants
maximum utilization of resources in efficient manner so that
its customers get satisfied with the services. To fulfil this
purpose concept of virtualization is implemented.
Virtualization requires virtual machine manager, which
manage the mapping of resource to virtual machines, creation,
deletion and migration of virtual machines from one system to
Fig 8 Type II Hypervisor [20] another. This feature of VMM makes the system extremely
robust and responsive in nature [24].
IV. ISSUES & CHALLENGES
5) Automated Service Provisioning: Because of the flexible
With the passage of time, cloud computing is getting more
feature of cloud-based system, resources can be assigned or
and more popular. Many small and large scale organizations
taken back any time. The problem associated with elastic
are using the cloud services. Besides the advantages of cloud,
property of cloud is to give the service as per the service level
the reliability of cloud services and data security are some of
agreement through the available group of resources. This
the major hurdles in wide acceptance of cloud. Important
results in a requirement of strategy which ensures optimum
research concerns in the field of cloud computing are energy
management of available resources for maintaining the same
management and data management, data security, optimum
level of performance as mentioned in the service level
resource utilization, multi-tenancy, automated service
agreement [24].
provisioning, charging model, virtual machine migration and
geographical variation. Their discussions are as follows: 6) Optimize Resource Utilization: This includes optimum use
of available resources thereby minimizing the consumption of
1) Data Security: As cloud is providing services through energy and reducing the complexity of implementation. This
internet, so along with cloud-specific security issues it also can be attained by load balancing. It includes resource
possess the internet security issues. Traditional cryptography allocation to virtual machine, virtual machine migration from
based security solutions are not sufficient enough to improve one node to another node and associating the task with virtual
the users trust in the cloud. There is need of some role-based machine. This issue of load balancing is an NP-hard
security mechanisms to ensure confidentiality, integrity, and optimization problem. Algorithms for load balancing should
authenticity [21]. be designed and implemented in such a manner that it
2) Multi-Tenancy: Multi-occupancy and openness are the optimally utilizes the resources without increasing the
crucial feature of cloud computing. Because of this, cloud complexity [24].
service provider is giving diverse services and facilities to the 7) Energy Management: 53% of total data center's working
end user. However, this has also resulted in some security charge occur because of cooling and power charge. The aim
issue. Those are as follows: of energy management is to keep the environment green by
o As the data and resources are stored at multiple minimizing the usage of energy [24].
locations in the cloud so during security infiltration
8) Data Management: Data sources like weather prediction
it becomes tough to trace the physical location.
system, social media sites, and geographical information
o Due to partnership of multiple organizations in
systems are generating structured and unstructured data in
cloud implementation, it becomes difficult to
exponential manner. This has become the reason of
implement single security policy.
emergence of big data concept. Traditional data models are
o Due to multi-tenancy and the open nature feature of
not able to handle big data. This has caused the need of new
cloud environment, different users share same
data model, which is capable of handling big data. [25].
resources at the same time. So some mechanism is
required to maintain privacy among the different V. CONCLUSION & FUTURE SCOPE
user [21].
Although cloud computing is in early phase of its
3) Charging Model: Shifting towards cloud has removed the development yet everyone wants to adopt it due to digitization.
startup cost barrier. However, on the other side, this has given There are some default challenges associated with cloud
birth to storage and communication cost. In addition, non- computing due to its basic characteristics and usage of internet.
transparency in charging model of cloud has given birth to These challenges are becoming a hurdle in wide adoption of
some more issues. The charging model is not transparent in cloud computing. In this paper, authors have analyzed the
nature [22]. The user has to pay according to the number & evolution of cloud, along with related terms and challenges.

IJSRCSAMS
Volume 7, Issue 4 (July 2018) www.ijsrcsams.com
 ISSN 2319 – 1953
International Journal of Scientific Research in Computer Science Applications and Management Studies

Research can be extended corresponding to every discussed [23] A. Keshavarzi, A.T. Haghighat, and M. Bohlouli, “Research challenges
and prospective business impacts of cloud computing: a survey”, In
challenge. Big data, security, cost model and load balancing
Proceedings of 7th International Conference on Intelligent Data
are some of the areas in which research is still going on and Acquisition and Advanced Computing Systems (IDAACS),Vol. 2, pp.
further can be extended. 731-736, 2013.
[24] A. Jain and R. Kumar, “Critical analysis of load balancing strategies
REFERENCES for cloud environment”, International Journal of Communication
Networks and Distributed Systems, Vol. 18, Issue 3-4, pp.213-234,
[1] C. Curricula, “The overview report. A volume of the Computing
2017.
Curricula Series”, A cooperative project of the ACM, the AIS, the
[25] S. Sharma, “An Extended Classification and Comparison of NoSQL
IEEE-CS, 2005.
Big Data Models”, arXiv preprint arXiv:1509.08035, 2015.
[2] G.S. Almasi, and A. Gottlieb, “Highly parallel computing”, 1988.
[3] G. Coulouris, J. Dollimore, and T. Kindberg,, “Distributed Systems
Concepts and Design”, Pearson Education, 2012.
[4] G. F. Pfister, “In search of clusters”, Prentice-Hall Inc., 1998.
[5] S. Zhang, X. Chen, S. Zhang, and X. Huo, “The comparison between
cloud computing and grid computing”, In the Proceedings of
International Conference on Computer Application and System
Modelling (ICCASM 2010), Vol. 11, pp. V11-72, 2010
[6] I. Foster, Y. Zhao, I. Raicu and S. Lu, “Cloud computing and grid
computing 360-degree compared”, In Proceedings of Grid Computing
Environments Workshop, pp. 1-10, 2008.
[7] R. Buyya, C. Vecchiola and S.T. Selvi, “Mastering cloud computing:
foundations and applications programming”, Newnes, 2013.
[8] A. Jain and R. Kumar, “A Taxonomy of Cloud Computing”,
International Journal of Scientific and Research Publications, Vol. 4,
Issue 7, pp.1-5, 2014.
[9] B. Sosinsky, “Cloud computing bible”, John Wiley & Sons, 2010.
[10] N. Sadashiv and S.D. Kumar, “Cluster, grid and cloud computing: A
detailed comparison”, In Proceedings of 6th International Conference
on Computer Science & Education (ICCSE), pp. 477-482, 2011.
[11] B. Yu, J. Tian, S. Ma, S. Yi and D. Yu, D., “Gird or cloud? Survey on
scientific computing infrastructure”, In Proceedings of 2011 IEEE
International Conference on Cloud Computing and Intelligence
Systems, pp. 244-249, 2011.
[12] P. Mell and T. Grance, “The NIST definition of cloud computing”,
2011.
[13] L. Wang, J. Tao, M. Kunze, A.C. Castellanos, D. Kramer and W. Karl,
“Scientific Cloud Computing: Early Definition and Experience”, .In
HPCC, Vol. 8, pp. 825-830, 2008.
[14] C. Vecchiola, S. Pandey and R. Buyya, “High-performance cloud
computing: A view of scientific applications”, In Proceedings of 10th
International Symposium on Pervasive Systems, Algorithms, and
Networks, pp. 4-16, 2009.
[15] C. Gong, J. Liu, Q. Zhang, H. Chen and Z. Gong, “The characteristics
of cloud computing”, In Proceedings of 39th International Conference
on Parallel Processing Workshops, pp. 275-279, 2010.
[16] L.J. Zhang and Q. Zhou, “CCOA: Cloud computing open architecture”,
In Proceedings of International Conference on Web Services, pp. 607-
616, 2009.
[17] X.J. Chen, B.D. Chen, X.M. Jiang, X.B. Chen, and W.H. Cai,
“Improved Cloud Computing Architecture for the Internet of Things”,
Journal of Internet Technology, Vol. 17, Issue 4, pp. 683-693, 2016.
[18] X. Wang, B. Wang and J. Huang, “Cloud computing and its key
techniques”, In Proceedings of International Conference on Computer
Science and Automation Engineering (CSAE), Vol. 2, pp. 404-410,
2011.
[19] Y. Xing and Y. Zhan, “Virtualization and cloud computing”, Future
Wireless Networks and Information Systems, pp. 305-312, Springer
Berlin Heidelberg, 2012.
[20] F. Lombardi and R. Di Pietro, “Secure virtualization for cloud
computing”, Journal of Network and Computer Applications, Vol. 34,
Issue 4, pp.1113-1122, 2011.
[21] A. Jain and R. Kumar, “Confidentiality Enhanced Security Model for
Cloud Environment”, 2nd ACM International Conference on
Information and Communication Technology for Competitive
Strategies, pp.139-145, 2016.
[22] A Vouk, “Cloud computing–issues, research and implementations”,
Journal of Computing and Information Technology, Vol. 16, Issue 4,
pp.235-246, 2008.

IJSRCSAMS
Volume 7, Issue 4 (July 2018) www.ijsrcsams.com