SCALABLE COMMUNICATION FOR WEB APPLICATIONS BY USING WINDOWS AZURE SERIVE

7/27/2019 SCALABLE COMMUNICATION FOR WEB APPLICATIONS BY USING WINDOWS AZURE SERIVE

1/4

International Journal of Computer Trends and Technology- volume4Issue3- 2013

ISSN: 2231-2803 http://www.internationaljournalssrg.org Page 316

SCALABLE COMMUNICATION FOR WEB APPLICATIONS BY

USING WINDOWS AZURE SERIVE1T.Geethika , 2Sk.Sunaina, 3 Jaladi.Vinusha

1,2,3 M.Tech Computer Science Students KLUniversity, Guntur, Andhra Pradesh, India

Abstract:

Cloud computing has emerged as aideal platform for deploy scalable web-applications. But at the same time theygive up data consistency. However, manyapplications cannot give any datainconsistency scalable transactionmanager which guarantee ACIDproperties for multi-item connectionsissued by Web application, even in theattendance of server failure and networkpartitions. We execute this approach ontop of the two main families of scalabledata layers: Bigtable and SimpleDB.Performance evaluation on top of HBase(an open-source version of Bigtable) inour local cluster and Windows AzureTable Storage Service in the Windowscloud shows that our system scaleslinearly at least up to 40 nodes in ourlocal cluster and 80 nodes in the Windowscloud.

Keywords: Scalable,cloud computing, Webapplications, Transactions

Introduction:

The extensive reputation of Cloud computing as a favored platform for theoperation of web applications has resulted inan huge number of applications affecting tothe cloud, and the enormous victory of cloud

service providers. Due to the rising number of web applications being hosted in the

cloud, and the increasing scale of data whichthese applications accumulate, procedure,and serve up scalable data managementsystem form a decisive part of cloud infrastructures .

.Figure 1: Typical Software Stack for Web

Applications

The scalability and high accessibility properties of Cloud platforms thoughapproach at a price. Initially, These scalabledatabase services allocate data query only by

primary key rather than sustainingsecondary-key or join queries. Next, theseservices give only feeble reliability such as


2/4



ultimate data consistency: any data keepinformed become visible after a limited butun deterministic amount of time. As feebleas this reliability possessions may seem, itdoes allow to build a wide range of usefulapplications, as established by the profitablesuccess of Cloud computing platforms.Though, a lot of additional applications suchas expense and online public sale servicescannot afford any data inconsistency. While

primary key- only data admittance is arelatively minor problem that can often beaccommodate by good data structures, it isnecessary to provide transactional dataconsistency to support the applications thatrequire it.

A transaction is a set of queries to beexecute atomically only on reliable view of a database. The main test to supporttransactional guarantee in a cloud computingatmosphere is to give the ACID propertiesof Atomicity, Consistency, Isolation and Durability with no compromise thescalability property of the cloud. Though,the underlying cloud data storage services

provide only ultimate reliability. We addressthis inconsistency by create a secondary

provisional copy of the application data inthe transaction managers that gripconsistency.

We exhibit the scalability of our transactional database service usingexecution. The data models of Bigtable and SimpleDB, communication are allowable toaccess any number of data items by primarykey at the granularity of the informationline. The listing of primary keys access byan operation must be known clearly beforeexecute the transaction. This means for instance that variety queries are not supportinside a transaction.

We appraise our trial product under aworkload derived from the TPC-W e-commerce benchmark. We implemented Cloud TPS on top of two different scalable

data layers: H Base, an open-source replicaof Big Table, organization in our localcluster; and Simple DB, organization in theWindows Cloud . We show that Cloud TPSscales linearly to at least 40 LTMs in our local cluster and 80 LTMs in theWindowsCloud. This income that, accordingto the values of Cloud computing, any raisein workload be able to be accommodate by

provisioning extra servers.

2 RELATED WORK:

2.1 Data Storage in th e Cloud

The simplest way to store ordered data in the cloud is to organize a relationaldatabase such as MySQL or Oracle. Therelational data model, classically implementvia the SQL language, provide hugeelasticity in accessing data. It supportscomplicated data access operation such asaggregation, range queries, join queries, etc.RDBMSs support communication and warranty strong data reliability. One can

simply deploy a classical RDBMS in thecloud and thus get sustain for transactionalreliability. Though, the elastic querylanguage and strapping data reliability avertone from partition data mechanically, whichis the key for presentation scalability. Thesesystem rely on duplication technique and therefore do not carry additional scalabilitydevelopment compare to a non-cloud operation.

On the other hand, a new family of

cloud database services such as WindowsAzure,Google Bigtable , AmazonSimpleDB, Yahoo PNUTS, and Cassandra,uses easy data models based on qualityvalue pair. Request data are organized intotables, of which each is a compilation of data items. Data items are typically accessed through a GET/ PUT interface by primarykey. Additional difficult data access


3/4



operations, such as range queries in SimpleDB or data item scanning in Big table, arelimited within a table. None of themsupports operation crosswise numeroustables, such as join queries. This data modelallow such systems to panel application datainto any number of tables efficiently.

2.2 Windows Azure Table StorageService :

The Windows Azure calculateservice can run many different kinds of applications. A main objective of this

platform, however, is to support applicationsthat have a very large number of concurrentusers. Reaching this goal by scaling up running on bigger and bigger machines isnt possible. Instead, Windows Azure is

planned to support applications that scaleout, running multiple copies of the samecode across many commodity servers.

To allow this, a Windows Azureappliance can have numerous instances,each executing in its own virtual machine(VM). These VMs run 64-bit WindowsServer 2008, and theyre provided by ahypervisor (based on Hyper-V) thats beenmodified for use in Microsofts cloud. Torun an application, A developer accesses theWindows Azure portal through her Web

browser, signing in with a Windows LiveID. She then chooses whether to create ahosting account for running applications, astorage account for storing data, or both.Once the developer has a hosting account,she can upload her application, specifyinghow many instances the application needs.Windows Azure then creates the necessaryVMs and runs the application.

Its important to note that adeveloper cant supply her own VM imagefor Windows Azure to run the platform itself

provides and maintains its own copy of Windows. Developers focus solely on

creating application that run on WindowsAzure.

The goal of Microsofts WindowsAzure is to provide this Part of the larger Azure Services Platform.Windows Azure isa platform for running Windowsapplications and storing data in the cloud.Figure 2 illustrates this idea.

Figure 2: Windows AzureApplication

3.ACID Properties:

Each of the property is discussseparately. We then talk about themembership mechanism to assurance theACID properties even in case of LTMfailure and network partition .

Atomicity

Atomicity refers to the ability of theDBMS to guarantee that either all of thetasks of a transaction are performed or noneof them are. Atomicity states that databasemodifications must follow an all or nothing rule. If some part of a transactionfails, then the entire transaction fails, and vice versa.Consistency

The Consistency property ensuresthat the database remains in a consistent


4/4



state, despite the transaction succeeding or failing and both before the start of thetransaction and after the transaction is over.Isolation

Isolation refers to the requirementthat other operations cannot access or seethe data in an intermediate state during atransaction. As discussed above, theIsolation property can helps to implementconcurrency of database.

Durability

Durability states that once atransaction is committed, its effects areguaranteed to persist even in the event of subsequent failures. That means when usersare notified of success, the transactions will

be persist, not be undone and survive fromsystem failure.

4.Conclusion:Web applications require strong data

consistency for their accurateimplementation. Even though the highscalability and ease of use property of thecloud make it a good platform to host Websatisfied, scalable cloud database servicessimply supply comparatively weak consistency property. This article showshow one can sustain strict ACIDtransactions with no compromise thescalability property of the cloud for Webapplications. Windows Azure provides largedatabase service to the providers and hostthese web applications these platform is for running windows applications and stores inthe cloud .

5.References :

[1]http://download.microsoft.com/download /e/4/3/e43bb484-3b52-4fa8-

a9f9ec60a32954bc/Azure_Services_Platfor m.pdf [2] B. Hayes, Cloud computing,Communications of the ACM , vol. 51,

no. 7, pp. 911, Jul. 2008[3] ABADI, D. J., MADDEN, S. R., ANDFERREIRA, M. C. Integrating compressionand execution in columnoriented databasesystems. Proc. of SIGMOD (2006).[4] BANGA, G., DRUSCHEL, P., ANDMOGUL, J. C. Resource containers: A newfacility for resource management in server systems. In Proc. of the 3rd OSDI (Feb.1999), pp. 4558.

Documents

SCALABLE COMMUNICATION FOR WEB APPLICATIONS BY USING WINDOWS AZURE SERIVE