Backup Resource Allocation Model with Two-Stage Probabilistic Protection

Abstract

This paper proposes a backup resource allocation model with two-stage probabilistic protection to minimize the total required backup capacity for multiple simultaneous failures of physical machines (PMs). Probabilistic protection ensures that the probability that the PM used for backup fails to backup due to lack of computing capacity does not exceed a given survivability parameter. The proposed model protects the primary virtual machines by allocating computing capacity to backup PMs with probabilistic protection. Since it is uncertain which primary PMs fail, we apply robust optimization to the probabilistic protection. By using a table that takes into account the survivability parameter and the failure probability of PMs, this proposed model is formulated as a mixed integer linear programming problem. The proposed model extends the probabilistic protection to two stages; the VMs that fail to be protected in the first stage are protected in the second stage to achieve the probabilistic protection with the final survivability parameter. This model can reduce the total required backup capacity compared to the conventional model with one-stage probabilistic protection.