Distributed Computing System: Cloud Computing Research Report

1
DISTRIBUTED COMPUTING SYSTEM
CLOUD COMPUTING
RAMAN PAL
([email protected])

M.Sc. Web Technology

School of Electronics & Computer Science
University of Southampton, UK

Abstract: Cloud Computing is one of the Area can be traced from the evolution of Grid
biggest buzz word in the present ICT scenario. Computing and Distributed Computing.
There is no perfectly coined definition for this Cloud Computing has initiated a new market
word which seems a bit poetic at times. of Services over the internet. And the key
However, it is referred as service which is players in this market are big shots IT
provided over the internet and scalability, Companies like: Amazon, Google, Microsoft,
virtualization its resources. Services in cloud IBM, etc.
computing are IaaS (Infrastructure as a Cloud Computing can be considered as a
Service), SaaS (Software as a Service), PaaS subset of the grid computing as they share
(Platform as a Service). This paper will throw same technology maintaining the key ethics of
some light over Cloud Computing, its Distributed Computing.
applications and the problems which are As this technology is now being considered as
associated with it. And will try to suggest one of the most important revolutionising
some solutions to obstacles it has in its way. technology of today’s ICT scenario, it also has
some obstacles lined up in its way to achieve
Keywords: Distributed Computing, Cloud the glory of becoming a super technology.
Computing, IaaS, SaaS, PaaS, etc. Obstacles and research problems like: Data
transfer bottlenecks, Reputation Fate Sharing,
1. Introduction
Software Licensing, Data Lock-In, Security,
Cloud computing is a specialized distributed Data Confidentiality and Performance
computing paradigm, which deals with Unpredictability. These research problems are
lending services on a pay-as-you-go basis over the major obstacles of Cloud Computing
the internet. As per [1] Cloud Computing is becoming a success.
“A large-scale distributed computing paradigm This paper will explain the Cloud Computing
that is driven by economies of scale, in which a technology its features, services, application,
pool of abstracted, virtualized, dynamically- will compare the contrast between it to its
scalable, managed computing power, storage, predecessors Grid Computing and Distributed
platforms, and services are delivered on demand to Computing, and will try to suggest some
external customers over the Internet.”[1] solutions to the obstacles that are lined up in
Cloud Computing will be one of the many front of Cloud Computing Success.
important milestones in opening the gates
toward the success of Web 3.0.
The roots of this new technology in
Information & Communication Technology
 Distributed Computing System: Cloud Computing Research Report

2. Cloud Computing Cloud Computing offers its services by

different means they are: 2
The dream of computing as a utility like:
Electricity, Water, Gas, etc. is on the verge of a) Infrastructure as a Service: IaaS refers to
becoming true with advent of Cloud on-demand provisioning of infrastructural
Computing as a revolutionising technology in resources, usually in terms of VMs. The cloud
today’s ICT scenario. Cloud computing as a owner who offers IaaS is called an IaaS
technology has immense potential of setting a provider. Examples of IaaS providers include
benchmark by making the Software more Amazon EC2, GoGrid and Flexiscale. [3,8]
attractive and lucrative as a service to its
b) Platform as a Service: PaaS refers to
buyers in the IT World.
providing platform layer resources, including
Defining Cloud Computing operating system support and software
There is no unified definition for Cloud development frame works. Examples of PaaS
Computing. Cloud Computing is considered as Providers include Google AppEngine,
a “Collection of scalable virtualized resources, Microsoft Windows Azure and Force.com.
which is capable of hosting application and [3,8]
providing required services to the users and
c) Software as a Service: SaaS refers to
can charge as per the uses like utility. It
providing on-demand applications over the
provides infrastructure as a service, software
Internet. Examples of SaaS providers include
as a service, platform as a service, network as
Salesforce.com, Rackspace and SAP Business
a service, and data storage as a service. The
by Design. [3,8]
main philosophy of cloud computing is to
provide every required things as a service.”[2] 3. Characteristic features of
In Cloud Computing the collective effort of Cloud Computing
entities like software, hardware and network a) Ultra Large Scale: The cloud is a
play the important role. It provides virtual largely scaled. Google’s cloud has
server infrastructures for companies. The owned more than one million servers.
intended benefit is that enterprises do not Amazon, IBM, Microsoft, Yahoo, they
have to buy their own hardware to provide have more than hundreds of
services for their customers. thousands servers. Hundreds of
servers are there in an enterprise.
Cloud computing enables developers with Cloud enlarges the user’s computing
innovative ideas for new Internet services no power.
longer require the large capital outlays in b) Virtualization: Cloud computing
hardware to deploy their service or the enables users to get service
human expense to operate it. They need not anywhere, anytime and through any
be concerned about over-provisioning for a kind of terminal. The required
service whose popularity does not meet their resources come from cloud instead of
predictions, thus wasting costly resources, or visible entity. You can complete all
under-provisioning for one that becomes you want through internet service
wildly popular, thus missing potential using a notebook, desktop terminal or
customers and revenue. even a PDA. Users can attain or share
it safely through an easy way,
anytime, anywhere.
 Distributed Computing System: Cloud Computing Research Report

c) High reliability: Data multi-transcript single technology was capable enough for the
fault tolerant system is used by smooth functioning of Cloud Computing. 3
Clouds, the computation node Technologies such Grid Computing,
isomorphism exchangeable and so on Distributed Systems, Virtualization, Web
to ensure the high reliability of the Services, Service Oriented Architecture,
service. Using local desktop terminal System Engineering, Utility Computing and
is less reliable than cloud computing. Autonomic Computing. Each of the above
d) Versatility: Cloud computing doesn’t technologies worked together to form a new
aim at certain special application. technology called Cloud Computing.
Cloud can produce various supported
applications and it can support Grid Computing: It is also a paradigm of
different applications running them at Distributed Computing System. A common
the same time. computational objective is achieved by
e) High extendibility: Dynamic coordinating networked resources. Scientific
extensibility can scale the cloud to applications which were computation
meet increasing requirement. intensive led to the development of Grid
f) On demand service: You can buy or Computing. Cloud computing is similar to Grid
pay on the go for resources you use computing in that it also employs distributed
from the large pool of resources resources to achieve application-level
called Cloud; cloud is just like running objectives. However, cloud computing takes
water, electric, and gas that can be one step further by leveraging virtualization
charged by the amount that you used. technologies at multiple levels (hardware and
g) Extremely inexpensive: Because the application platform) to realize resource
cloud’s special fault tolerance can be sharing and dynamic resource provisioning.
built by very inexpensive nodes, the Virtualization: Virtualization is a technology
centered management of cloud make that abstracts away the details of physical
the enterprise needn’t undertake the hardware and provides virtualized resources
management cost of data center that for high-level applications. A virtualized server
increase very fast. The versatility can is commonly called a virtual machine (VM).
increase the utilization rate of the Virtualization forms the foundation of cloud
available resources compared with computing, as it provides the capability of
traditional system, so users can fully pooling computing resources from clusters of
enjoy the low cost advantage. You can servers and dynamically assigning or
spend only a few hundred dollars and reassigning virtual resources to applications
a few days to accomplish a task that on-demand.[3]
you must do it spending thousands of
dollars and several months before. Utility Computing: Utility computing
represents the model of providing resources
on-demand and charging customers based on
usage rather than a flat rate. Cloud computing
4. Technologies in Cloud can be perceived as a realization of utility
Computing computing. It adopts a utility-based pricing
scheme entirely for economic reasons. With
Various technologies were combined and
on-demand resource provisioning and utility-
integrated to form the cloud computing. No
based pricing, service providers can truly
 Distributed Computing System: Cloud Computing Research Report

maximize resource utilization and minimize intervention. The goal of autonomic

their operating costs. [3] computing is to overcome the management 4
complexity of today’s computer systems.
Autonomic Computing: Originally coined by Although cloud computing exhibits certain
IBM in 2001, autonomic computing aims at
autonomic features such as automatic
building computing systems capable of self- resource provisioning, its objective is to lower
management, i.e. reacting to internal and the resource cost rather than to reduce
external observations without human
system complexity. [3]
5. Cloud Computing Architecture

Figure 1: Cloud Computing Architecture [3]

6. Comparing different commercial products for Cloud Computing[3]

 Distributed Computing System: Cloud Computing Research Report

7. Cloud Computing services offered by different service providers[4]

5

8. Research Problems and advancement in the field of Cloud Computing

there are issues and research problems which
Obstacles
are still struggling for a definite solution.
Cloud Computing is relatively new technology
and new developments and researches are a) Data Transfer Bottle Necks:
taking place every day. Even after so much Applications continue to become
 Distributed Computing System: Cloud Computing Research Report

more data-intensive. If we assume performance than in main memory

applications may be “pulled apart” performance. 6
across the boundaries of clouds, this
may complicate data placement and Solutions/Suggestion: Better & improved
transport. At $100 to $150 per Virtual Machines Support. Regular
terabyte transferred, these costs can memories and I/O disks should be
quickly add up, making data transfer replaced with high performance flash
memories and solid State Drives. Regular
costs an important issue.
checks and logs should be maintained.
Solution/Suggestion: They should deploy
novel data management approaches, such d) Software Licensing: Many cloud
as analytical data management tasks, computing providers originally relied
multitenant databases for SaaS, or hybrid on open source software in part,
designs among database management because the licensing model for
systems (DBMSs) and MapReduce-like commercial software is not a good
systems so as to address data limitations match to utility computing.
and harness cloud computing platforms’ Solution/Suggestion: Software, in a cloud,
capabilities [7]. Or a combination of becomes a service. To a company like
system where manual conventional Microsoft, whose substantial fortunes are
methods like FedEx and Advance technical built on software as a purchasable local
solutions like online sharing/transfer application [9]. New licensing model
should be implemented, depending on should be built and implemented, which
the circumstances and needs of the will allow users and consumers to use the
system the correct methodology should software on pay-as-you-go basis. Or new
be implemented. developments in open source software
b) Reputation Fate Sharing: One should be encouraged in school and
customer’s bad behaviour can affect colleges this will result in more and better
the reputation of others using the software under open source hood.
same cloud. e) Security, Data Confidentiality: Data
Solution/Suggestion: Reputation safe management by the third party
guarding mechanisms should be service provider has always been a
implemented and regular quality issue of concern.
assurance testing should be done. Solutions/Suggestion: Significant amounts
Feedbacks and User reviews should be of data which were previously stored only
regarded and considered important. in individual offices and homes would
c) Performance Un-predictability: now reside in data centers controlled by
Studies revealed that multiple virtual third-parties [9]. Better and improved
machines (VMs) can share CPU’s and security mechanisms specially tailored for
main memory surprisingly well in the cloud computing application should
cloud computing, but that network be developed. Better encryption
and disk I/O sharing is more methodologies and Virtual private LANs
problematic. As a result, different EC2 and Firewalls should be used.
instances vary more in their I/O
 Distributed Computing System: Cloud Computing Research Report

f) Business Continuity and Service images, and so on, with a variety of

Availability: Organizations worry cloud computing providers and 7
about whether utility computing platforms. Cloud interoperability will
services will have adequate enable cloud infrastructures to evolve
availability, and this makes some wary into a worldwide, transparent
of cloud computing. platform in which applications aren’t
restricted to enterprise clouds and
Solutions/Suggestion: Use of multiple
cloud service providers. We must
cloud computing providers and there build new standards and interfaces
services. Integration among different that will enable enhanced portability
Cloud Computing services should be
and flexibility of virtualized
encouraged and implemented. applications.
g) Data Lock-In: Concern about the 9. Major concern for cloud
difficulty of extracting data from the computing: Security
cloud is preventing some Security is one of the largest concerns for the
organizations from adopting cloud adoption of Cloud Computing. We outline
computing. seven risks a Cloud user should raise with
vendors before committing [6]:
Solution/Suggestion: API’s should be
standardized and use of Hybrid a) Privileged user access: sensitive data
Computing Cloud should be encouraged processed outside the enterprise needs the
over the user private computing cloud. assurance that they are only accessible and
New compatible SW should be developed propagated to privileged users;
to enable Surge Control.
b) Regulatory compliance: a customer needs
h) Scalable Storage: The opportunity, to verify if a Cloud provider has external udits
which is still an open research and security certifications and if their
problem, is to create a storage system infrastructure complies with some regulatory
that would not only meet existing security requirements;
programmer expectations in regard to
durability, high availability, and the c) Data location: since a customer will not
ability to manage and query data, but know where her data will be stored, it is
combine them with the cloud important that the Cloud provider commit to
advantages of scaling arbitrarily up storing and processing data in specific
and down on demand. jurisdictions and to obey local privacy
requirements on behalf of the customer;
Solution/Suggestion: New and advanced
storage solutions should be developed which d) Data segregation: one needs to ensure
will be capable of Scaling up and down as per that one customer’s data is fully segregated
requirement, without developing any flaw in from another customer’s data;
its functionality.
e) Recovery: it is important that the Cloud
i) Cloud Interoperability [7]: Cloud provider has an efficient replication and
interoperability refers to customers’ recovery mechanism to restore data if a
ability to use the same artifacts, such disaster occurs;
as management tools, virtual server
 Distributed Computing System: Cloud Computing Research Report

f) Investigative support: Cloud services are [3] “Cloud computing: state-of-the- art and
especially difficult to investigate, if this is research challenges” Q. Zhang, L. Cheng, R. 8
important for a customer, then such support Boutaba
needs to be ensured with a contractual
[4] “Cloud Computing–A Classification,
commitment;
Business Models, and Research Directions”
g) Long-term viability: your data should be Prof. Dr. C. Weinhardt, Dr. B. Blau, Dr. J.
viable even the Cloud provider is acquired by Stößer
another company.
[5] “A View of Cloud Computing” M.
10. Conclusion: ArmbruSt, A. Fox, R. Griffith, A. D. Joseph, R.
Cloud computing is an emerging technology in Katz, A. KonWinski, G. LEE, D. Patterson, A.
which every services are available in the Rabkin, I. Stoica, M. Zaharia
cloud. Cloud is the collection of distributed
[6] “Gartner: Seven cloud-computing security
computing devices. Cloud computing will
risks”, J. Brodkin.
grow with time, so developers should take it
http://www.networkworld.com/news/2008/0
into account. Scalable Software and pay-as-
70208-cloud.html, 2008.
you licensing model should be considered.
Security policies should be revised and new [7] “Cloud Computing: Distributed Internet
policies and trust models should be Computing for IT and Scientifc Research” M.
developed. Many key challenges in this D. Dikaiakos, G. Pallis, D. Katsaros, P. Mehra,
domain, including automatic resource A. Vakali. IEEE Computer Society 2009
provisioning, power management and security
management, are only starting to receive [8] “Semantic Computing, Cloud Computing,
attention from the research community. and Semantic Search Engine”
There is still tremendous opportunity for P. C-Y SHEU, S. WANG, Q. WANG, K. HAO, R.
researchers to make ground breaking PAUL. 2009 IEEE International Conference on
contributions in this field of study. This paper Semantic Computing
presented the state-of-the-art of cloud
computing, covering its essential concepts, [9] “COMPUTING IN THE CLOUDS” A.Weiss
architectural designs, prominent
characteristics, key technologies, comparison
between service providers as well as
discussed and presented many research
challenges and obstacles.

References:
[1] “Cloud Computing and Grid Computing
360-Degree Compared” I. Foster, Y. Zhao, I.
Raicu, S. Lu.

[2] “Cloud Computing: Future solution for e-

Governance” M. Pokharel, J. S. Park