A Change-Based Three-Way Decision Model for Virtual Machine Migration and Consolidation Optimization

Given the swift advancement of cloud computing technology, virtual machine migration has become an important component of data center resource scheduling. However, existing migration trigger strategies are often insufficiently precise, leading to suboptimal virtual machine consolidation effects. Simultaneously, symmetric load fluctuations in high and low ranges also pose challenges for energy-efficient scheduling. Therefore, this paper proposes an energy-efficient virtual machine migration and consolidation method based on change-based three-way decision (EEVM-C3WD). This method first calculates the load fluctuation of the physical machine, and then divides it into three areas: high, medium, and low through threshold division. Subsequently, the C-3WD mechanism is introduced to reconstruct these three regions based on current load and confidence level changes, thereby obtaining three finer-grained regions. For different reconstructed regions, different prediction models are employed, and the upper and lower thresholds are dynamically adjusted accordingly to trigger migration more precisely. Experimental results show that by appropriately increasing the number of migrations, EEVM-C3WD significantly improves the number of physical machines that can be shut down and the virtual machine consolidation success rate, ultimately achieving the goal of energy saving and consumption reduction.