Вы находитесь на странице: 1из 4

Cloud Computing: State Of The Art Reseach Issues

Babak Abbasov
Department of Information Technology and Systems
Qafqaz University
Baku, Azerbaijan
babbasov@qu.edu.az

Abstract— Cloud computing has emerged as a new desktops, entertainment centers, table computers,
technology aimed to host computer resources and deliver notebooks, wall computers, handhelds, etc.
MIT Technology …the idea of relying on Web-based application
appropriate services over the Internet. Cloud computing and storing data in the “CLOUD” of the internet.
become attractive to IT industry by providing user Eric Schmidt It starts with the premise that the data services and
requirements in pay as you go principle and increase (Chairman/CEO of architecture should be on the servers. We call it
resources only when service demand increases. However, Google Inc) Cloud Computing – they should be in a “CLOUD”
somewhere.
despite the fact that cloud computing offers tremendous Garter “A style of computing in which massively scalable
opportunities to the IT industry, there are still many issues ITrelated capabilities are provided “as a service”
still to be addressed in the development of cloud using Internet technologies to multiple external
computing technology. In this paper, a review of cloud customers”[1].
IDC “An emerging IT development, deployment and
computing technology and its concepts, state-of-the-art delivery model, enabling realtime delivery of
implementation, brief architecture and research challenges products, services and solutions over the Internet
have been investigated. (i.e., enabling cloud services)” [2].
The 451 Group “A service model that combines a general
Keywords: cloud computing; services; SaaS; cloud resource organizing principle for IT delivery, infrastructure
management components, an architectural approach and an
economic model – basically, a confluence of grid
computing, virtualization, utility computing,
I. INTRODUCTION hosting and software as a service (SaaS)” (Fellows
2008)
Cloud computing has become a tremendous paradigm for
hosting and delivering services over the Internet. The goal of
this computing model is to make a better use of various Five essential characteristics of cloud computing is clearly
computer resources, keep together in order to achieve higher articulated in these definitions:
throughput and to resolve problems require high performance
computations. It allows users to develop and leverage On-demand self-service. A user can provision cloud
available computer resources in pay as you go manner. resources, such as server time and network storage,
automatically without requiring human interaction with each
service provider.
II. CLOUD: OVERVIEW
Broad network access. Cloud resources are available over
A. Definitions of Cloud Computing the internet and accessed through diverse terminals such as
Diverse formal definitions have been proposed in both thin or thick client platforms (e.g., computers, mobile phones,
research studies and IT industry. Widely used definitions in the tablets etc.).
Cloud Computing community are illustrated in Table 1. Resource pooling. The computing resources are pooled
together to serve multiple consumers in a multi-tenant manner.
TABLE I. CLOUD C OMPUTING DEFINITIONS These physical and virtual resources are assigned as the
consumer needed. Customer generally has no control or
Name Definition knowledge over the exact location of the provided computer
Nist Cloud computing is a model for enabling
ubiquitous, convenient, on-demand network access resources. But customer may be able to specify location at a
to a shared pool of configurable computing country, state, or datacenter etc. levels. Computer resources
resources (e.g., networks, servers, storage, provided to customers include storage, processing, memory,
applications, and services) that can be rapidly and network bandwidth.
provisioned and released with minimal
management effort or service provider interaction. Rapid elasticity. Computer resources provisioned to
This cloud model is composed of five essential customers can be elastically provisioned and released and can
characteristics, three service models, and four be scaled rapidly outward and inward appropriately with
deployment models [3].
IEEE It is a paradigm in which information is
demand. Consumer generally is able to scale out resources that
permanently stored in servers on the Internet and appears to be unlimited and can be provisioned in any quantity
cached temporarily on clients that include at any time.
Measured service. Computer resources can be optimize the utilization of existing in-house resources. Second,
automatically controlled and optimized using a metering security concerns including data privacy and trust also make
capability appropriate to the type of service provisioned (e.g., Private Cloud an option for many firms. Third, data transfer
storage, processing, bandwidth, and active user accounts). cost [2] from local IT infrastructure to a Public Cloud is still
Resource utilization can be monitored, controlled, and rather considerable. Fourth, organizations always require full
reported for both the provider and consumer to provide control over mission-critical activities that reside behind their
transparency. firewalls. Last, academics often build private cloud for
research and teaching purposes.
B. Cloud Services Community cloud. Several organizations jointly construct
Following cloud service models have been used to classify and share the same cloud infrastructure as well as policies,
the cloud services in cloud community: requirements, values, and concerns. The cloud community
Software as a Service (SaaS). Cloud user release their forms into a degree of economic scalability and democratic
applications on a hosting environment, which can be accessed equilibrium. The cloud infrastructure could be hosted by a
through internet from diverse terminals (e.g. web browser, third-party vendor or within one of the organizations in the
PDA, etc.). Cloud consumers do not have control over the a community.
multi-tenant cloud infrastructure. Cloud consumers' Public cloud. This is the dominant form of current Cloud
applications are managed in a single virtual environment on computing deployment model. The public cloud is used by the
the SaaS to leverage optimized amount of resources in terms general public cloud consumers and the cloud service provider
of availability, speed, security, maintenance and disaster has the full ownership of the public cloud with its own policy,
recovery. Examples of SaaS include SalesForce.com, Google value, and profit, costing, and charging model. Many popular
Mail, Google Docs, Amazon etc. cloud services are public clouds including Amazon EC2, S3,
Platform as a Service (PaaS). PaaS is a development Google AppEngine, and Force.com.
platform supporting the full "Software Lifecycle" which Hybrid cloud. The cloud infrastructure is a combination of
allows cloud consumers to develop cloud services and two or more clouds (private, community, or public) that
applications (e.g. SaaS) directly on the PaaS cloud. Hence the remain unique entities but are bound together by standardized
difference between SaaS and PaaS is that SaaS only hosts or proprietary technology that enables data and application
completed cloud applications whereas PaaS offers a portability (e.g., cloud bursting for load-balancing between
development platform that hosts both completed and in- clouds). Organizations use the hybrid cloud model in order to
progress cloud applications. optimize their resources to increase their core competencies by
This requires PaaS, in addition to supporting application margining out peripheral business functions onto the cloud
hosting environment, to possess development infrastructure while controlling core activities on-premise through private
including programming environment, tools, configuration cloud. Hybrid cloud has raised the issues of standardization
management, and so forth. An example of PaaS is Google and cloud interoperability that will be discussed in later
AppEngine. sections.The template is used to format your paper and style
the text. All margins, column widths, line spaces, and text
Infrastructure as a Service (IaaS). Cloud consumers fonts are prescribed; please do not alter them. You may note
directly use IT infrastructures (processing, storage, networks, peculiarities. For example, the head margin in this template
and other fundamental computing resources) provided in the measures proportionately more than is customary. This
IaaS cloud. Virtualization is extensively used in IaaS cloud in measurement and others are deliberate, using specifications
order to integrate/decompose physical resources in an ad-hoc that anticipate your paper as one part of the entire proceedings,
manner to meet growing or shrinking resource demand from and not as an independent document. Please do not revise any
cloud consumers. The basic strategy of virtualization is to set of the current designations.
up independent virtual machines (VM) that are isolated from
both the underlying hardware and other VMs. Notice that this D. A layered model of cloud computing
strategy is different from the multi-tenancy model, which aims Cloud computing architecture can be divided into 4 layers:
to transform the application software architecture so that
the hardware datacenter layer, the infrastructure layer, the
multiple instances (from multiple cloud consumers) can run on platform layer and the application layer, as shown in Table. 2.
a single application (i.e. the same logic machine). An example
Detailed description of these layers is as follows:
of IaaS is Amazon's EC2.

C. Deployment Models TABLE II. A LAYERED MODEL OF CLOUD COMPUTING

More recently, four cloud deployment models have been Cloud computing Resource Example applications
defined in the Cloud community: services management
Applications, web Facebook,
Software as a Service
Private cloud. The cloud infrastructure is operated solely (SaaS)
services, multimedia Youtube,
within a single organization, and managed by the organization Application Google Apps
or a third party regardless whether it is located premise or off Platform as a Service Software framework Google AppEngine,
(PaaS) (Java, .Net), Storage Microsoft Azure,
premise. The motivation to setup a private cloud within an Amazon Simple DB/S3
Platform
organization has several aspects. First, to maximize and
Computation, Storage Amazon EC2, overcoming lack of limited resources, resource heterogeneity,
Infrastructure as a Flexiscale locality restrictions, environmental necessities and dynamic
Infrastructure
Service (IaaS) CPU, Memory, Disk Data Centers nature of resource demand.
Hardware
Finding an optimum resource allocation strategy in huge
systems like data centers is expansive due to dynamic and
The hardware layer refers to managing the physical uncertain resource demand and supply. Various resource
resources of the cloud computing infrastructure, including allocation schemes are proposed in practice.
physical servers, routers, switches, power and cooling
systems. This layer is implemented in data centers and B. Cloud Security, Privacy and Reliability
contains thousands of servers which arranged in racks and Important concerns about cloud computing are security
interconnected through switches, routers or other techniques. and privacy. Company confidential data must be unavailable
The infrastructure layer (the virtualization layer) manages and inaccessible by other companies. Related research issues
a pool of storage and computing resources by partitioning the on resource allocation techniques done in the literature
physical resources using virtualization techniques (Xen [16], [11:15].
KVM [17] and VMware [18]). Companies offering cloud computing services must satisfy
security concerns. Customers only pay these companies when
The platform layer includes operating systems and they are convinced in security measures. Contrarily companies
application frameworks. The aim of this layer is to minimize advance security concerns to not lose their customers.
the burden of deploying applications directly into virtual Privacy is another issue. As user’s data can be accessed
machine containers. from any location, proper authentication techniques should be
The application layer consists of the cloud applications developed. When using authorization each user can access
which can be exploited with the automatic-scaling feature to only the data and applications relevant to his or her job.
achieve better performance, availability and lower operating Replication time and costs also play an important role due
cost. to how fast can the data be replicated is important for data
resiliency.
Reliability is an issue. Server’s stability should be
III. STATE OF THE ART RESEACH ISSUES
provided to avoid downtimes that can occur with cloud
Recently, researchers focus on diverse areas of cloud servers.
computing regarding to the importance. In this paper state of
the art research issues investigated on the cloud computing. C. Virtual machine migration
Virtualization technique in cloud computing can be used to
A. Cloud Resource Management increase efficiency, balance load across the data centers and
enables robust and highly responsive provisioning in data
Resource Allocation is the process of provisioning
available computer resources to satisfy request made from centers.
cloud applications. Applications require type and amount of
resources in order to complete a user process. The order and D. Server consolidation
time of allocation of resources are also important. Related Server consolidation is an approach to maximize resource
research issues on resource allocation techniques done in the utilization while minimizing energy consumption in a cloud
literature [4:10]. A resource allocation schema should avoid computing. Some under-utilized servers can be set to an
the following criteria: energy-saving state by combine VMs located on several
servers onto a single server using Virtual machine migration.
• Resource contention arises when two applications
want to access the same resource at the same time.
E. Energy management
• Scarcity of resources arises when there are limited Increasing energy efficiency in cloud computing and
resources. designing energy-efficient data centers has recently received
• Resource fragmentation situation arises when the considerable attention. Energy efficient hardware architecture
resources are isolated. There will be enough resources but not that enables slowing down CPU speeds and turning off partial
able to allocate to the needed application. hardware components [19], energy-aware job scheduling [20]
and server consolidation [21] used to reduce power
• Over-provisioning of resources arises when the consumption by turning off unused machines, studying
application gets surplus resources than the demanded one. energy-efficient network protocols and infrastructures [22] are
• Under-provisioning of resources occurs when the all intended to achieve a good trade-off between energy
application is assigned with fewer numbers of resources than savings and application performance.
the demand.
F. Traffic management and analysis
For a cloud provider, predicting the dynamic nature of
Analysis of network traffic is important for today’s data
cloud users, user and application demands are inapplicable.
centers to increase efficiency. Network operators also need to
For the cloud users, the task should be completed on time with
know how traffic flows through the network in order to make
minimal cost. An efficient resource allocation scheme requires
many of the management and planning decisions. Many Informatization (ICSEM), 2011 International Conference on , vol.2,
techniques used to increase network performance in computer pp.233,236, 22-23 Oct. 2011
networks [23, 24, 25]. Also, it is important to improve [11] Almorsy, M.; Grundy, John; Ibrahim, AS., "Collaboration-Based Cloud
Computing Security Management Framework," Cloud Computing
network performance in datacenter networks. (CLOUD), 2011 IEEE International Conference on, pp.364,371, 4-9 July
2011
IV. CONCLUSION [12] Demchenko, Y.; Canh Ngo; De Laat, C.; Wlodarczyk, T.W.; Chunming
Rong; Ziegler, W., "Security Infrastructure for On-demand Provisioned
In this paper a new paradigm in the field of information Cloud Infrastructure Services," Cloud Computing Technology and
technology: cloud computing is discussed. Cloud computing Science (CloudCom), 2011 IEEE Third International Conference on ,
architecture, advantages and some issues are described. There pp.255,263, Nov. 29 2011-Dec. 1 2011
is no doubt that cloud computing is the development trend for [13] Feng Zhao; Chao Li; Chun Feng Liu, "A cloud computing security
the future. solution based on fully homomorphic encryption," Advanced
Communication Technology (ICACT), 2014 16th International
Conference on , pp.485,488, 16-19 Feb. 2014
References [14] Chalse, R.; Selokar, A; Katara, A, "A New Technique of Data Integrity
for Analysis of the Cloud Computing Security," Computational
[1] Gartner (2008b) Gartner Says Contrasting Views on Cloud Computing Intelligence and Communication Networks (CICN), 2013 5th
Are Creating Confusion. Gartner press release, 29 September 2008. International Conference on , vol., no., pp.469,473, 27-29 Sept. 2013
[2] Gens F (2008) Defining “Cloud Services” and “Cloud Computing”. IDC [15] Wentao Liu, "Research on cloud computing security problem and
eXchange, 23 September 2008. strategy," Consumer Electronics, Communications and Networks
[3] Mell, Peter, Grance, Timothy, “The NIST Definition of Cloud (CECNet), 2012 2nd International Conference on, pp.1216,1219, 21-23
Computing”, National Institute of Standards and Technology (NIST) April 2012
(Hrsg.), Gaithersburg, MD, pp. 800-145, 2011. [16] XenSource Inc, Xen, www.xensource.com
[4] Weinong Wu; Xing Zhang; Yuanbing Zheng; Huili Liang, "Agent-based [17] Kernal Based Virtual Machine, www.linux-kvm.org/page/MainPage
layered cloud resource management model," Information Management, [18] VMWare ESX Server, www.vmware.com/products/esx
Innovation Management and Industrial Engineering (ICIII), 2013 6th
International Conference on, pp.70,74, 23-24 Nov. 2013 [19] D.M. Brooks et al, “Power-Aware Microarchitecture: Design and
Modeling Challenges for Next-Generation Microprocessors”, IEEE
[5] Grant, AB.; Eluwole, O.T., "Cloud resource management — Virtual Micro, 20(6):36-44, Nov./Dec. 2000.
machines competing for limited resources," AFRICON, 2013, pp.1,7, 9-
12 Sept. 2013 [20] Nedeljko Vasić, Martin Barisits, Vincent Salzgeber, and Dejan Kostic,
"Making cluster applications energy-aware", In Proceedings of the 1st
[6] Tran, G.S.; Broto, L.; Hagimont, D., "Toward Informed Resource workshop on Automated control for datacenters and clouds (ACDC '09),
Management in the Cloud," Utility and Cloud Computing (UCC), 2013 ACM, New York, NY, USA, 37-42, 2009
IEEE/ACM 6th International Conference on, pp.247,250, 9-12 Dec.
2013 [21] S. Srikantaiah, A. Kansal, and F. Zhao, “Energy aware consolidation for
cloud computing,” in Proc. of HotPower, 2008
[7] Jie Bao; Zhihui Lu; Jie Wu; Shiyong Zhang; Yiping Zhong,
"Implementing a novel load-aware auto scale scheme for private cloud [22] IEEE P802.3az Energy Efficient Ethernet Task Force,
resource management platform," Network Operations and Management www.ieee802.org/3/az
Symposium (NOMS), 2014 IEEE, pp.1,4, 5-9 May 2014 [23] Abbasov, B., "AHRED: A robust AQM algorithm for wireless ad hoc
[8] Dabbagh, M.; Hamdaoui, B.; Guizani, M.; Rayes, A, "Energy-efficient networks," Application of Information and Communication
cloud resource management," Computer Communications Workshops Technologies, 2009. AICT 2009. International Conference on , vol., no.,
(INFOCOM WKSHPS), 2014 IEEE Conference on, pp.386,391, April pp.1,4, 14-16 Oct. 2009
27 2014-May 2 2014 [24] Abbasov, B., "Using linear interpolation method for packet drop
[9] Xing Chen; Ying Zhang; Xiaodong Zhang; Xianghan Zheng; Wenzhong probability function of RED algorithm," Application of Information and
Guo; Guolong Chen, "Architecture-Based Integrated Management of Communication Technologies (AICT), 2011 5th International
Cloud Resources," Cloud Computing and Big Data (CloudCom-Asia), Conference on , vol., no., pp.1,4, 12-14 Oct. 2011
2013 International Conference on , vol., no., pp.474,481, 16-19 Dec. [25] Abbasov, B., Korukoglu S., "Effective RED: An algorithm to improve
2013 RED's performance by reducing packet loss rate", Journal of Network
[10] Yuan Yuan; Wen-Cai Liu, "Efficient resource management for cloud and Computer Applications, Volume 32, Issue 3, May 2009, Pages 703-
computing," System Science, Engineering Design and Manufacturing 709

Вам также может понравиться