Fast Communication-Aware Virtual Machine Dynamic Consolidation for Cloud Data Center

With help of virtualization technology, cloud data centers provide affordable computation power to numerous enterprises and business users. Virtual machine dynamic consolidation is considered as an approach that can effectively save energy consumption of cloud data centers and improve its run time efficiency. However, most existing approaches perform virtual machine dynamic consolidation without considering communication traffic between VMs. Such kind of approaches may lead to unsatisfactory consolidation result because inappropriate consolidation may exacerbate inter-rack traffic, making upper level switches performance bottlenecks for the entire system. In this paper, we propose a fast communication-aware consolidation approach based on a distributed architecture. We modify SCAN algorithm to discover VM communication topology. The most attractive benefit of SCAN is its low computation cost. According to such communication topology information, we propose a dynamic consolidation heuristic that can mitigate inter-rack traffic. Through extensive simulations, we show effectiveness of our approach and its advantage over state of arts in terms of energy saving, average traffic through core switches and task response time.