Certain investigations on performance analysis of data replication in cloud computing for data as a service DaaS

Abstract

Cloud computing is a computation through internet which shares the resources and services to users on demand The cost of infrastructure is reduced perhaps computing power and storage capacity is increased Though the amount of structured and unstructured data are increasing day by day due to social media scientific research and web based applications the cloud storage services are highly recommended for these types of data which can be accessed anywhere at any time Because of huge data the data storage accomplished in multiple data nodes in which the probability of failure is high This rises unreliable connection and limited bandwidth in the network In order to increase data availability reliability and consistency between data has become the prime challenge in the cloud storage Data replication has been used widely as a means to increase data availability and reliability in the cloud storage system But only by increasing the number of replicas do not always increase data availability instead increases storage space and cost which is not beneficial for both users and cloud providers In recent years dynamic replication plays a vital role in the distributed computing field which create replicas with respect to the change in access patterns and location of the replica which changed dynamically on the storage of capacity of nodes Further dynamic replication has three important issues: Firstly when is data file to be replicated? Secondly finding optimal number of new replicas for meeting the system availability and reliability with minimum cost of replica maintenance Thirdly placement of replicas to meet the successful execution rate and BW consumption requirement In addition consistency among the replicas are to be maintained newline