Principles of transaction-oriented database recovery
ACM Computing Surveys (CSUR)
Performance analysis of recovery techniques
ACM Transactions on Database Systems (TODS)
On the Optimum Checkpoint Interval
Journal of the ACM (JACM)
Hi-index | 0.98 |
This paper examines the reliability of the database recovery mechanism by using archive copies for fault latency. An evaluation model for the recovery is proposed to derive the recovery overhead for a latent fault and the availability for a cycle. The results of numerical calculations show the impacts of model parameters considered. It is concluded that, for the total system availability in normal situations, an archive recovery scheme should be determined taking account of only nonlatent faults and the maintenance costs of multiple copy generations, whether the effect of fault latency is great or not. On the other hand, if the system puts restrictions on the recovery time, the archive copy interval and the number of copy generations should be considered to determine the recovery scheme.