International Journal of Scientific & Engineering Research, Volume 5, Issue 3, March-2014 1084
ISSN 2229-5518
k.Sindhuja, P.Anand Raj
Abstract— Cloud-based hosting is claimed to possess many advantages over traditional in-house hosting such as better scalability, ease of management, and cost savings. It is not difficult to understand how cloud-based hosting can be used to address some of the existing limita- tions and extend the capabilities of many types of applications. There are several algorithms already used for resource allocation in cloud computing. Based on analyzing these existing cloud resource allocation schemes, the new resource allocation algorithm based on hybrid fuzzy clustering will be introduced. The proposed algorithm cleverly assigns appropriate resource to the tasks that exactly satisfy its need for resource, while reserving the resources whose power greatly exceed the needs of current tasks for future use. By this way, the proposed algorithm can effectively avoid assigning powerful resources to simple and medium scale tasks or assigning poor resources to complex large-scale tasks, for they may lead to misuse of resources and failure in scheduling of tasks. The proposed algorithm is expected to have high efficiency and good robustness and will work better than other similar algorithms. The highlight of a new many-to-many stable matching algorithm that efficiently matches virtual machines with heterogeneous resource needs to servers. This work will provide a new approach for cloud based computing.
Index Terms— Cloud computing, Resource Scheduling, Hybrid Fuzzy Clustering, Virtual Machine Placement, Stable Matching.
—————————— ——————————
loud computing is the delivery of computing services over the Internet. Cloudservices allow individuals and businesses
to use software and hardware that aremanaged by third par- ties at remote locations. The cloud computing model allows access to information and computerresources from anywhere that a network connection is available. Cloud computingpro- vides a shared pool of resources, including data storage space, networks,computer processing power, and specialized corpo- rate and user applications.
Cloud computing has a variety of characteristics, with the main ones being:
based on current demand requirements. This is done automat- icallyusing software automation, enabling the expansion and
————————————————
• Anand Raj Assistant Professor, E.G.S.Pillay Engineering college, Nagapat-
tinam. E-mail: anand.happy123@gmail.com
contraction of service capability, as needed. This dynamic scal- ingneeds to be done while maintaining high levels of reliabil- ity and security.
mizing the service and to provide reporting and billing infor- mation.In this way, consumers are billed for services according to how much they have actually used during the billing peri- od.In short, cloud computing allows for the sharing and scala- ble deployment of services, as needed, from almost any loca- tion, andfor which the customer can be billed based on actual usage.
Once a cloud is established, how its cloud computing services are deployed in terms of business models can differ depend- ing onrequirements. The primary service models being de- ployed.
Consumers purchase the ability to access and use an applica- tion or service that is hostedin the cloud. A benchmark exam- ple of this is Salesforce.com, as discussed previously, where necessary information for theinteraction between the consum- er and the service is hosted as part of the service in the cloud.
IJSER © 2014 http://www.ijser.org
The research paper published by IJSER journal is about Efficient Framework For Resource Management In Cloud Using Hybrid Fuzzy Clustering
ISSN 2229-5518
Also, Microsoft is expanding its involvement in this area, and
1085
as part of the cloud computing option for Microsoft® Office
2010, itsOffice Web Apps are available to Office volume licens-
ing customers and Office Web App subscriptions through its cloud-basedOnline Services.
Consumers purchase access to the platforms, enabling them to deploy their own software andapplications in the cloud. The operating systems and network access are not managed by the consumer, and there might beconstraints as to which applica- tions can be deployed.
Consumers control and manage the systems in terms of the operating systems,applications, storage, and network connec- tivity, but do not themselves control the cloud infrastruc- ture.Also known are the various subsets of these models that may be related to a particular industry or market. Communi- cations as aService (CaaS) is one such subset model used to describe hosted IP telephony services. Along with the move to CaaSis a shift tomore IP-centric communications and more SIP trunking deployments. With IP and SIP in place, it can be as easy to have the PBXin the cloud as it is to have it on the prem- ise. In this context, CaaS could be seen as a subset of SaaS.
Deploying cloud computing can differ depending on require- ments, and the following four deployment models have been identified,each with specific characteristics that support the needs of the services and users of the clouds in particular ways.
and is maintained and operated for a specific organization. Theoperation may be in-house or with a third party on the premises.
lic on a commercial basis by a cloud service provider. Thisena- bles a consumer to develop and deploy a service in the cloud with very little financial outlay compared to the capitalex- penditure requirements normally associated with other de- ployment options.
of clouds of any type, but the clouds have the ability throughtheir interfaces to allow data and/or applications to be moved from one cloud to another. This can be a combination of privateand public clouds that support the requirement to retain
some data in an organization, and also the need to offer ser- vices inthe cloud.
The following are some of the possible benefits for those who offer cloud computing-based services and applications:
tures and use operational expenditures for increasing theircomputing capabilities. This is a lower barrier to entry and also requires fewer in-house IT resources to provide sys- tem support.
ployment and grow to a large deployment fairly rapidly, and thenscale back if necessary. Also, the flexibility of cloud com- puting allows companies to use extra resources at peak times, enablingthem to satisfy consumer demands.
port business continuity and disaster recovery.
nance, and access is through APIs that do not require applica- tioninstallations onto PCs, thus further reducing maintenance requirements.
ity due to systems accessible in an infrastructure available fromanywhere.
The following are some of the notable challenges associated with cloud computing, and although some of these may cause aslowdown when delivering more services in the cloud, most also can provide opportunities, if resolved with due care and attentionin the planning stages.
issues surrounding cloud computing relate to storing andse- curing data, and monitoring the use of the cloud by the service providers. These issues are generally attributed to slowingthe deployment of cloud services. These challenges can be ad- dressed, for example, by storing the information internal to theorganization, but allowing it to be used in the cloud. For this to occur, though, the security mechanisms between organ- izationand the cloud need to be robust and a Hybrid cloud could support such a deployment.
however, no standards are associated with these, and thus it isunlikely that most clouds will be interoperable. The Open Cloud Forum is developing an Open Cloud Computing Inter- face toresolve this issue and the Open Cloud Consortium is working on cloud computing standards and practices. The findings ofthese groups will need to mature, but it is not known whether they will address the needs of the people de- ploying the servicesand the specific interfaces these services need. However, keeping up to date on the latest standards as they evolve will allowthem to be leveraged, if applicable.
IJSER © 2014 http://www.ijser.org
The research paper published by IJSER journal is about Efficient Framework For Resource Management In Cloud Using Hybrid Fuzzy Clustering
ISSN 2229-5518
1086
evolving, as are the requirements for interfaces, networking, andstorage. This means that a “cloud,” especially a public one, does not remain static and is also continuously evolving. Compliance Concerns - The Sarbanes-Oxley Act (SOX) in the US and Data Protection directives in the EU are just two amongmany compliance issues affecting cloud computing, based on the type of data and application for which the cloud is beingused. The EU has a legislative backing for data protec- tion across all member states, but in the US data protection is differentand can vary from state to state. As with security and privacy mentioned previously, these typically result in Hybrid clouddeployment with one cloud storing the data internal to the organization.
There has been various types of scheduling algorithm exist in distributed computing system. Most of them can be applied in the cloud environment with suitable verifications. The main advantage of job scheduling algorithm is to achieve a high performance computing and the best system throughput. Tra- ditional job scheduling algorithms are not able to provide scheduling in the cloud environments. According to a simple classification, job scheduling algorithms in cloud computing can be categorized into two main groups; Batch mode heuris- tic scheduling algorithms (BMHA) and online mode heuristic algorithms. In BMHA, Jobs are queued and collected into a set when they arrive in the system. The scheduling algorithm will start after a fixed period of time. The main examples of BMHA based algorithms are; First Come First Served scheduling algo- rithm (FCFS), Round Robin scheduling algorithm (RR), Min– Min algorithm and Max–Min algorithm by On-line mode heu- ristic scheduling algorithm, Jobs are scheduled when they ar- rive in the system. Since the cloud environment is a heteroge- neous system and the speed of each processor varies quickly, the on-line mode heuristic scheduling algorithms are more appropriate for a cloud environment. Most fit task scheduling algorithm (MFTF) is suitable example of On-line mode heuris- tic scheduling algorithm.
come first is served. This algorithm is simple and fast.
cesses are dispatched in a FIFO manner but are given a limited amount of CPU time called a time-slice or a quantum. If a pro- cess does not complete before its CPU-time expires, the CPU is preempted and given to the next process waiting in a queue. The preempted process is then placed at the back of the ready list.
executed firstly, which in turn large task delays for long time.
executed firstly, which in turn small task delays for long time. Most fit task scheduling algorithm: In this algorithm task which fit best in queue are executed first. This algorithm has high failure ratio.
ward: each process is assigned a priority, and priority is al- lowed to run. Equal-Priority processes are scheduled in FCFS order. The shortest-Job-First (SJF) algorithm is a special case of general priority scheduling algorithm. An SJF algorithm is simply a priority algorithm where the priority is the inverse of the (predicted) next CPU burst. That is,the longer the CPU burst, the lower the priority and vice versa. Priority can be defined either internally or externally. Internally defined pri- orities use some measurable quantities or qualities to compute priority of a process.
Anchor is a general resource management architecture thatus- es the stable matching framework to decouple policies[1] from mechanisms when mapping virtual machines to physical serv- ers.
In Anchor, clients and operators are able to express a variety of distinct resource management policies as they deem fit and these policies are captured as preferences in the stable match- ing framework.
The highlight of Anchor is a new many-to-one stable matching
theory that efficiently matches VMs with heterogeneous re- source needs to servers, using both offline and online algo- rithms. Our theoretical analysis shows the convergence and optimality of the algorithm.
Fig 1. Anchor Architecture
1. A Resource Monitor
2. A Policy Manager
3. A Matching Engine
IJSER © 2014 http://www.ijser.org
The research paper published by IJSER journal is about Efficient Framework For Resource Management In Cloud Using Hybrid Fuzzy Clustering
ISSN 2229-5518
1087
• Both the operator and its clients are able to configure theirresource management policies, based on per- formance, cost, etc., as they deem fit via the policy manager.
• When VM placement requests arrive, the policy man- ager polls information from the resource monitorand feeds it with the policies to the matching engine.
• The matching mechanism resolves conflicts of interest among stakeholders, and outputs a matching be- tween VMs and servers.
• It is based on the stable matching framework from economics,which elegantly achieves all the design objectives.
• Specifically, the concept of preferences is used to ena- ble stakeholders to express various policies with simple rank-ordered lists,fulfilling the requirement of generality and expressiveness.
• Rather than optimality, stability is used as the central
solution concept to address the conflicts of interest among stakeholders, fulfilling the fairness require- ment.
• Finally, its algorithmic implementations based on the
classical Deferred Acceptance algorithm (DA) have
been demonstrated to bepractical in many real-world applications fulfilling theefficiency requirement.
• Jobs propose to machines following the order in their preferences.
• We randomly pick any free job that has not pro- posed to every machine on its preference to pro- pose to its favourite machine that has not yet re- jected it.
• That machine accepts the most favourable offers
made so far up to the capacity, and rejects the rest.
• Individual rationality needs to be respected for the matching to be acceptable to all participants.
• Optimization solvers are computationally expen-
sive due to their combinatorial nature, and do not scale well.
• Different size VMS.
• Size heterogeneity make easy .
• Job-optimal weakly stable matching, in the
sense that every job is assigned its best machine possible in all stable matching’s.
• A Revised DA Algorithm guaranteed to find a weakly stable matching for a given problem.
• Multistage DA algorithm to iteratively find a
better weakly stable matching with respect to jobs and it produces a correctness, convergence, and job optimality.
• Many-to-one stable matching theory that effi- ciently matches VMs with heterogeneous re- source needs to servers.
• Many-to-one problem where one server can en-
roll multiple VMs but one VM can only be as- signed to one server. Preferences are used as an abstraction of policies no matter how they are de- fined.
• Uniform VMs so virtual machine stable to work in demand of resource so speed of the resource management decreases.
• Efficiency also decreases.
• Two problems arise when applying the classical
DA algorithm here.
• First, the execution sequence is no longer imma-
terial to the outcome.
• Second, it may even produce an unstable match- ing. This creates considerable difficulties since we cannot determine which proposing sequence yields a weakly stable matching for an arbitrary problem.
• The DA algorithm fails precisely due to the size heterogeneity of jobs.
The proposed algorithm assigns appropriate resource to the very task that exactly satisfy its needs for resource, while re- serving the resources whose power greatly exceed the needs of current tasks for future use. By this way, our algorithm can effectively avoid assigning powerful resources to simple and medium scale tasks or assigning poor resources to complex large-scaletasks, for they may lead to misuse of resources and failure scheduling of tasks. The presented algorithm has high efficiency and good robustness and works better than other similar algorithms. Our work has provided a new approach for Cloud resource allocation.
In this proposed work,
By using Hybrid Fuzzy clustering algorithm[13] approach.A
new way to schedule appropriate resources to relevant tasks.It exactly satisfy the needs of tasks.Our algorithm avoid misuse of resources or failure scheduling of tasks when assigning po werful resources to simple or medium scale tasks or assigning poor resources to complex large scale tasks.
IJSER © 2014 http://www.ijser.org
The research paper published by IJSER journal is about Efficient Framework For Resource Management In Cloud Using Hybrid Fuzzy Clustering
ISSN 2229-5518
1088
Before allocating Cloud resources to Cloud tasks, firstly use a set of parameters.
Parameters are,
• CPU main frequency (fcpu )
• CPU free rate (CPUfree )
• Total memory size (MEMall )
• Available memory size (MEMavl )
• Available network bandwidth (NBavl )
• Available disk space (DISKavl )
So, The components of resource vector are
• R=(fcpu ,CPUfree ,MEMall ,MEMavl ,NBavl ,DISKavl ) The same way, the task vectors are
• T=(fcpu_by_task ,CPUfree_by_task ,MEMall_by_task ,MEMavl_by_task ,NBavl_
by _task , DISKavl_by_task )
Fig 2.Hybrid Fuzzy Clustering
• Resource table updation
• Analyzing the Task
• Mapping the task &Resource
• Adding the resource details in the database.
• Number of Resources of available in the Network- Cloud for the Application.
• Address the Resources.
• Find the suitable resource for the task.
• Combine the similar task in to the Group.
• Name the Groups and arrange in order.
• Using the Hybrid Fuzzy Clustering Algorithm map the resource with task.
• Our algorithm avoid misuse of resources or failure
scheduling of tasks when assigning powerful re- sources to simple or medium scale tasks or assigning poor resources to complex large scale tasks.
Fig 3. System Model
We investigate the performance of Hybrid Fuzzy Clustering with implementation.cloud sim is a web application that runs in a Virtual Machine on the Amazon webservice cloud.Fuzzy logic is a form of many valued logic.fuzzy logic variable may have truth value that ranges in degree between 0&1.Hybrid fuzzy clustering used to schedule the partially freed re- sources.The goal of fuzzy clustering is to determine the best grouping in a set of unlabeled data.The group or cluster n no of objects in to m number of groups depends on defined crite- ria. Hybrid fuzzy clustering is a new way to schedule appro- priate resources to relevant tasks.
IJSER © 2014 http://www.ijser.org
The research paper published by IJSER journal is about Efficient Framework For Resource Management In Cloud Using Hybrid Fuzzy Clustering
ISSN 2229-5518
1089
We achieve the good resource scheduling by using the Hybrid fuzzy clustering techniques. Basically it depends upon the concept of resource scheduling.
This work is related to research in the following fields.
An Optimal Model for Priority based Service Scheduling Poli-
cy for Cloud Computing Environment. Another planning cal-
culation dependent upon necessity and affirmation control
plan [2]. In this calculation necessity is doled out to each one conceded queue. Concession of each one queue is chosen by computing middle of the road postpone and administration cost.
Enhanced Cost-Based Algorithm for Task Scheduling. An en-
hanced expense based planning calculation for making pro- ductive mapping of assignments to accessible assets in cloud [8]. The act of spontaneity of accepted movement based setting back the ol' finances is proposed by new errand planning pro- cedure for cloud environment where there may be no connec- tion between the overhead provision base and the way that distinctive undertakings cause overhead cost of assets in cloud. This planning calculation separates all client undertak- ings relying upon necessity of each one assignment into three separate records. This planning calculation measures both asset expense and reckoning execution, it additionally im- proves computation/communication degree.
Stable matching [1]. Calculation to fathom the coupled posi-
tion of Vms in server farms. Our late work backer stable
matching as a general structure to tackle organizing issues. As
far as anyone is concerned, all these works accept an accepted
unisize employment model, while we concentrate on a more general size-heterogeneous model.
Our work is likewise identified with the written works on as- set booking.
The proposed technique of resource scheduling using hybrid fuzzy clustering is a new way to schedule appropriate re- sources to relevant tasks and it exactly satisfies the need of tasks. The presented algorithm has high efficiency and good robustness and work better than other similar algorithms. Our work has provided a new approach for cloud resource alloca- tion.
I would like to sincere thanks to Mr. Anand Raj, my project guide who gave me valuable advice and were good enough to find time for fruitful discussions.
[1] Hong Xu, Student Member, IEEE, and Baochun Li
,”Anchor: A Versatile and Efficient Framework for Resource Management in the Cloud”,ieee transactions on parallel and distributed systems, 2013, vol. 24, no. 6
[2] Dakshayini.M and Guruprasad H.S, “An Optimal Model
for Priority based Service Scheduling Policy for Cloud Com-
puting Environment”, IJCSEA (0975 – 8887) 2011Volume 32– No.9
[3] El-Sayed El-kenawy.T, Ali Ibraheem El-Desoky, Mohamed Al-rahamawy.F, “Extended Max-Min Scheduling Using Petri Net and Load Balancing”,2012, (IJSCE) ISSN: 2231-2307, Vol- ume-2.
[4] SaeedParsa and Reza Entezari-Maleki ”RASA: A New Task
Scheduling Algorithm in Grid Environment”, in World Ap- plied Sciences Journal 7 (Special Issue of Computer & IT), 1995 pp. 573-595. [5]ShalmaliAmbike,DiptiBhansali,JaeeKshirsagrandJuhiBansi wl “An Optimistic Differentiated Job Scheduling System for Cloud Computing”, (IJERA) ISSN: 2248-9622 www.ijera.com,
2012, Vol. 2, Issue 2, pp.1212-1214 .
[6] ShamsollahGhanbari and Mohamed Othman “A Priority
based Job Scheduling Algorithm in Cloud Computing”,
2012,International Conference on Advances Science and Con-
temporary Engineering (ICASCE).
[7] Selvarani1.S and sudha Sadhasivam.G”Improved cost-
based algorithm for task scheduling in Cloud computing”,
2010, IEEE .
[8] A.E. Roth “Deferred Acceptance Algorithms: History,
Theory,Practice, and Open Questions,” 2008, Int’l J. Game
Theory,vol. 36,pp. 537-569.
[9] B. Sotomayor, R. Montero, I. Llorente, and I. Foster “Virtu-
al Infrastructure Management in Private and Hybrid Clouds,”
2009 IEEE Internet Computing,vol. 13, no. 5, pp. 14-22.
[10] A. Kalbasi, D. Krishnamurthy, J. Rolia, and M. Richter
“MODE:Mix Driven On-Line Resource Demand Estimation,”
2011 Proc. IEEESeventh Int’l Conf. Network and Service Man- agement (CNSM).
[11] Z. Shen, S. Subbiah, X. Gu, and J. Wilkes, “CloudScale:
Elastic Resource Scaling for Multi-Tenant Cloud Systems,”
2011,Proc. ACM Second Symp. cloud Computing (SoCC) .
[12] S. Nascimento, B. Mirkin and F. Moura-Pires, “Multiple prototypes model for fuzzy clustering”, 1999 in J. Kok, D. Hand and M. Berthold, (eds.), Advances in Intelligent Data Analysis Third International Symposium,(IDA’99), Lecture Notes in Computer Science, 1642, Springer-Verlag, 1999,pp.
269–279.
[13] R. Dave, “Fuzzy shell-clustering and applications to circle
detection of digital images”, International Journal of General
Systems, 16(4), 1990 ,pp. 343–355,.
IJSER © 2014 http://www.ijser.org