A mathematical programming approach for relocation planning of data centers

This paper addresses the problem of relocation planning of data centers using mathematical optimization. In the literature, there are various problems in data centers that have been extensively studied using mathematical modeling; including Server Consolidation Problem (SCP) where virtual servers are required to be relocated to improve resource usage. To the best of the authors' knowledge, the problem of relocation planning of data centers, where servers are required to be physically relocated with minimum total cost associated with transfer time, has not been studied in the literature using analytical approaches based on mathematical models. Therefore, this paper proposes a one-dimensional Variable-Sized Bin Packing Problem (VS-BPP) formulation for this problem, which will be referred to as Server Relocation Problem (SRP). Then, an extended integer programming model for the SRP is proposed to consider several technical and administrative constraints that were found in a real-world case study from healthcare sector. The numerical results show that the data center considered in this paper can be relocated within 19 days of the given 50-day horizon with minimum total cost associated with transfer days while satisfying all practical requirements. Furthermore, this research suggests that analytical approaches can provide effective solutions for the SRP problem.