data replication in cloud computing

13
QoS-Aware Data Replication for Data-Intensive Applications in Cloud Computing Systems Guided by:- Ms.Shikha Jain Computer Science & Engineering JIIT NOIDA Submitted By:- Rahul Garg(10103499)

Upload: rahul-garg

Post on 30-Jun-2015

136 views

Category:

Education


2 download

TRANSCRIPT

Page 1: Data Replication In Cloud Computing

QoS-Aware Data Replication for Data-Intensive Applications in Cloud

Computing Systems

Guided by:-

Ms.Shikha JainComputer Science & EngineeringJIIT NOIDA

Submitted By:-Rahul Garg(10103499)

Page 2: Data Replication In Cloud Computing

Cloud Computing• An expression used to describe a variety of computing concepts.

• Involves a large number of computers connected through a real-time communication network such as the Internet.

• Relies on sharing computing resources rather than having local servers or personal devices to handle applications.

Page 3: Data Replication In Cloud Computing

Why Cloud Computing is important ?

• In traditional desktop computing we run software in our own PC and data is stored in there only, hence traditional computing is PC centric.

• Cloud computing is an emerging Paradigm in computer industry where

the computing is moved to a cloud of computers.

• Now documents are stored on a collection of servers accessed via internet. Hence the cloud computing is document centric.

Page 4: Data Replication In Cloud Computing
Page 5: Data Replication In Cloud Computing

Properties of Cloud Computing

• Powerful : Powerful in the sense that as there is large number of computers , thus more computing power and mass data storage possible.

• Task Centric : Cloud computing focus on what one need and how application can do it for us. Here documents are given more priority than the applications which create them.

• Flexible : Flexible as the users may be of different varieties and hence it has to match with their needs.

• Self Healing : Self healing because backups are available for every document in the cloud. Hence if one document crashes there will be it's duplicate ready to run.

Page 6: Data Replication In Cloud Computing

Problem Statement

In this project , we are focusing on “Quality of Service Aware Data Replication for Data Intensive Applications in Cloud Computing” which investigates the QADR problem for data-intensive applications in cloud computing systems.

QADR concerns how to efficiently consider the QoS requirements of applications in the data replication.

The main goal of the QADR problem is to minimize the data replication cost and the number of QoS violated data replicas.

Page 7: Data Replication In Cloud Computing

Data Intensive Applications Data-intensive applications are increasingly more demanding in

sophisticated algorithms to store, manage, and interpret data.

Storage and Security of such large data requires additional resources and innovative techniques.

Arising Problems

Page 8: Data Replication In Cloud Computing

Quality of Service

Quality of service is the overall performance of a computer network, particularly the performance seen by the users of the network

Quality of service is particularly important for the transport of traffic with special requirements.

Quality of service comprises requirements on all the aspects of a connection, such as service response time, signal-to-noise ratio, loss, interrupts, frequency response.

Page 9: Data Replication In Cloud Computing

Data Replication

Involves sharing of information so as to ensure consistency between redundant resources, such as software or hardware components.

To improve reliability, fault-tolerance, and accessibility.

Copies a database and synchronizes a set of replicas such that changes made to one replica are reflected in all the others.

Page 10: Data Replication In Cloud Computing

Mathematical Model

Where ,• y(ri,qj) is 1 if qj is a QoS violated replica,0 otherwise

• x(ri,qj) is 1 if qj is a qualified node for replication,0 otherwise

• Tstorage(ri,qj) is the total time for replication including Tdisk_access and Tnode_comm

Page 11: Data Replication In Cloud Computing

AlgorithmThe Goal is to minimize the two parameters :• Total replication cost in a qualified node• Number of QoS-violated data nodes

Input: A set of requested nodes Sr

Output: Optimal Placement for the QoS-satisfied and QoS-violated data replicas.

1. Sq NULL2. for each requested node ri in Sr do3. Sq

ri Find the correspondingly qualified nodes of ri

4. Sq Sq U Sq

ri

5. end for6. Use Sr and Sq

to model a network flow graph.

Page 12: Data Replication In Cloud Computing

7. Set appropriate (capacity, cost) values on the edges of the edges of the network flow graph.

8. Apply an existing polynomial-time MCMF algorithm to obtain the MCMF solution of the network flow diagram.

9. Obtain the optimal placement for the QoS-satisfied data replicas from the MCMF solution.

10. Sur NULL and Suq NULL11. For each requested node ri in Sr do12. fleaving the amount of flow leaving from ri

13. If fleaving < rf then

14. Sur Sur U ri 15. Sq

ri S - Sq

ri / * the qualified nodes for ri */

16. Suq Suq U Sqri

17. end if18. end for19. Use Sur and Suq to model a new network flow graph.20. Set appropriate (capacity, cost) values on the edges of the new network flow graph.21. Apply the MCMF algorithm on the new network flow graph to obtain the MCMF

solution.22. Follow the MCMF solution to make the optimal placement of the QoS- Violated data

replicas

Page 13: Data Replication In Cloud Computing