A METHOD FOR DETECTING THE CHANGED FILES IN A CLOUD-BASED BACKUP SYSTEM

Abstract

The cloud backup method for large amounts of data is presented which optimizes the change detection phase of incremental backup process. The incremental backup process opti- mization method is based on the dependency detection and analysis in large amounts of data, particularly, in the source code of websites. The cloud backup process can be organized with two different approaches. The first one is the full backup. The resource consumption, while performing full backup can be significant, particularly, when the amount of data in the system is large. Also, large full backups should be uploaded to remote backup servers, which is also resource-consuming. To overcome these disadvantages, the incremental backup approach is used in the Cloud Backup System. The advantage of incremental backup is that the only data that have been changed after the last backup are backed up, which means that the size of back- up archive is smaller and it will take fewer resources to create and upload it to the destination backup server. As a result, the problem of detection of changed files arises in the first phase of incremental backup process. The change detection is usually done by creating check values of data in the system and comparing them to the newly computed values while processing the backup. It is obvious, that if the number of files in the system is relatively large, then, the pro- cess of change detection will take significant amount of resources. The method of solving this issue is presented, which is based on dependencies between the files that are being backed up in the system.

The oriented dependency graph construction and the analysis algorithm are presented, the output of which is the list of files from the data that requires change detection.