The performance of applications running on the cloud entirely depends on two factors, namely, network availability and resource management. Resource contention occurs when request for resources to a host exceeds the availability of the resources and this leads to severe performance degradation of the application. Although virtualization has reduced the performance overhead, performance loss is still possible due to resource contention between collocated virtual machines (VMs). We propose a Second Order Markov Model based Prediction of Future State of Host algorithm for predicting resource contention in hosts in the cloud and to decide on the migration of VMs from one host to another. We also propose a Contention Mitigated Placement algorithm for placing the VMs that are migrated. The main objective of our work is to predict the hosts that will contend for resources and maximize the CPU utilization by reducing the number of VM migrations. Based on the predictions, the VMs from overloaded hosts are migrated to either under loaded or normally loaded hosts such that the destination host does not become overloaded after VM migration. As VM migration from one machine to another causes latency and decrease in CPU utilization due to migration overhead, we have used the number of VM migrations as a metric to measure the performance of our proposed work. Experimental results show that the proposed algorithms improve performance by reducing the number of VM migrations.

运行在云上的应用程序的性能完全取决于两个因素，即网络可用性和资源管理。当对主机的资源请求超过资源的可用性时，就会发生资源争用，这会导致应用程序的性能严重下降。尽管虚拟化降低了性能开销，但由于并置虚拟机 (VM) 之间的资源争用，性能仍有可能损失。我们提出了一种基于二阶马尔科夫模型的主机未来状态预测算法，用于预测云中主机的资源争用，并决定虚拟机从一台主机迁移到另一台主机。我们还提出了一种竞争缓解放置算法，用于放置已迁移的 VM。我们工作的主要目标是预测将竞争资源的主机，并通过减少 VM 迁移次数来最大化 CPU 利用率。根据预测，来自过载主机的 VM 将迁移到负载不足或负载正常的主机，以便在 VM 迁移后目标主机不会过载。由于从一台机器迁移到另一台机器会导致延迟和 CPU 使用率的降低，因为迁移开销，我们使用 VM 迁移的数量作为衡量我们建议工作性能的指标。实验结果表明，所提出的算法通过减少虚拟机迁移次数来提高性能。将来自过载主机的 VM 迁移到负载不足或负载正常的主机，以便目标主机在 VM 迁移后不会过载。由于从一台机器迁移到另一台机器会导致延迟和 CPU 使用率的降低，因为迁移开销，我们使用 VM 迁移的数量作为衡量我们建议工作性能的指标。实验结果表明，所提出的算法通过减少虚拟机迁移次数来提高性能。将来自过载主机的 VM 迁移到负载不足或负载正常的主机，以便目标主机在 VM 迁移后不会过载。由于从一台机器迁移到另一台机器会导致延迟和 CPU 使用率的降低，因为迁移开销，我们使用 VM 迁移的数量作为衡量我们建议工作性能的指标。实验结果表明，所提出的算法通过减少虚拟机迁移次数来提高性能。