Reliable and Available Data Replication Planning for Cloud Storage

Cloud storage (CS) is a type of data storage system widely accepted in information technology communities. Users of CS require reliable and continuous access to data on their stored data storage. Therefore, this study presents a data backup method for data retrieval and access from distinct places based on the access frequencies of a shared folder. The proposed method improves the quality of CS services by studying the data duplication of a CS system. This approach aims to minimize the data access time, which is subject to the required transmission bandwidth and location of data storage. Therefore, this study develops a network model, access frequency model, delay model, and budget to develop a new duplication mechanism. This study proposes a bar center method and evaluates its performance to determine how well the predicted CS access time outperforms the related methods.

I. Introduction

Information technology has become a significant factor in everyday life and business management. In the past, data backup involved copying data from a hard disk to offline record media, such as magnetic tapes and RAID (redundant array of independent disks) storage [1]. However, users must prepare personal computers and networks to install and configure backup servers. If the storage mediums fail or are lost, the data are lost. The capacity of offline storage is also limited [1], and the cost of maintenance and collection is high. Online storage is another type of backup offered by various web services (e.g., Sugar Sync, Dropbox, IBackup, MozyHome, and Memopal) [2]. This type of service gives users a remote access mechanism for collecting, compressing, encrypting, and transferring data to a backup system. However, users wanting to access data from an online storage must send a request to an administrator, which several users perceive as slow data access [3].