Improved Mathematical Model for Virtual Machine Placement Optimization with Resource Constraints

Virtual Machine placement problem is a hard-combinatorial optimization problem in cloud data center. In this paper we have described problem in detail with its importance and challenges. We have also proved that the problem is NP complete when multiple dimensional resources are considered during placement. To optimize such a complex problem, we have developed new MILP mathematical model by including various constrains. An empirical evaluation of this model has surpassed existing state-of-the-art MILP and branch and bound models and shown evaluation on different data sets have shown zero percentage gaps for more than 70% instances.