| The importance of data backup has been well understood in the IT industry. Computer systems are prone to failures/disruptions that result in data loss or corruption. A variety of causes can lead to the loss of electronic data. Examples of such causes include disk failures, human errors, virus infection, theft, and physical damage to equipment. Businesses as well as individuals often have important information such as financial records and business contacts kept in computers. This information is often made secure and available by the use of file level backup/restore systems. Once data is lost or corrupted, the only viable option left is to restore it from the backup media taken before the loss or corruption. Not taking any backups may lead to permanent loss of data.At present, most file-level backup system can not restore successfully without manual intervention. There are many reasons behind the disappointment. One of such reasons is the inability of these commercial backup/restore solutions to recognize the name or path changes on files. Among the remote backup and recovery systems, the known backup model: full backup, incremental backup and differential backup all have their own advantages and disadvantages. How could we take the best use of these three backup modes to improve restore efficiency and decrease the network transportation at the same time? How could we make our backup and recovery system independent of other software because of their log storage?To above three problems, this paper presents solutions for each of them. This paper proposes an Object ID-based File Match Scheme to addresses the file mismatch issue in traditional data backup solutions for NTFS environment. And an incremental differential backup mode and a log management method based on xml file. |
PDF Full Text Request