Heuristics based server consolidation with residual resource defragmentation in cloud data centers

• We model residual resource fragmentation in cloud data centers.
• We examine feasibility of residual resource defragmentation with server consolidation.
• Proposed approach performs defragmentation with low energy cost and SLA violations.
• Controlled defragmentation with consolidation reduces VM migrations.
• A good mix of cloud applications ensures better defragmentation.

Server Consolidation is one of the foremost concerns associated with the effective management of a Cloud Data Center as it has the potential to accomplish significant reduction in the overall cost and energy consumption. Most of the existing works on Server Consolidation have focused only on reducing the number of active physical servers (PMs) using Virtual Machine (VM) Live Migration. But, along with reducing the number of active PMs, if a consolidation approach reduces residual resource fragmentation, the residual resources can be efficiently used for new VM allocations, or VM reallocations, and some future migrations can also be reduced. None of the existing works have explicitly focused on reducing residual resource fragmentation along with reducing the number of active PMs to the best of our knowledge. We propose RFAware Server Consolidation, a heuristics based server consolidation approach which performs residual resource defragmentation along with reducing the number of active PMs in cloud data centers.