摘要

Cloud computing platforms often consist of a large number of physical computer nodes. With the increasing number of physical computer nodes, cloud computing platforms become more and more complex. Increasing complexity could make cloud computing platforms be prone to faults. In order to protect cloud computing platforms from failure, a hybrid push protocol is proposed for resource monitoring in cloud computing platforms in this paper. The proposed hybrid push protocol is designed by improving the basic push protocol and combines the advantages of the periodic push protocol and event-driven push protocol. It adopts the exponential weighted moving average to update the dynamic threshold and also introduces the lower limit and upper limit to protect the dynamic time interval from being too big or too small. A series of experiments have been conducted and experimental results show that our proposed hybrid push protocol outperforms the state-of-the-art data delivery protocols in terms of the data coherence and communication overhead.