An Evaluation of Restoration Time and the Amount of Backup Data of Data Replication Method for Large-scale Disasters
Hitoshi Kamei, Takaki Nakamura, Hiroaki MURAOKA · IEEJ Transactions on Electronics Information and Systems · 2017
Data replication method reducing data restoration time is evaluated on backup server fault model assuming Great East Japan Earthquake. Use-case is to back up important data via network in preparation to large-scale disasters. When the number of backup servers increases, restoration time becomes short because of parallel restoration; therefore, the restored data can be accessed as soon as possible. However, the amount of backup data also increases. In this study, data replication methods are classified, and evaluation functions are created for each method. Restoration time is calculated by using the backup server fault model, and then balanced method is revealed.