Data Backup in Geo-Distributed Data Center Networks Under Time and Budget Constraints
Lisheng Ma, Bin Yang · 2019
Region failure caused by disaster (e.g., earthquakes, hurricanes, etc.) can lead to the damages of geo-distributed data center networks (DCNs), which induce huge amounts of data loss. To implement data protection under disaster, data backup problem is considered in geo-distributed DCNs before DCNs are affected by an upcoming disaster and then an emergency data backup scheme is investigated, which takes the time constraint and budget differentiation into account. The former is used to ensure that data backup is carried out under the constraint of early warning time of disaster, while the latter meets the different budget constraints for data backup from different DCNs. To solve the data backup problem, an Integer Linear Program (ILP) is designed to obtain the optimal solution for backup under the time and budget constraints, such that the overall amount of backup data is maximized. The proposed ILP model is illustrated by extensive numerical results.