一种基于隐马尔可夫风险评估模型的数据快速疏散策略

A Fast Data Evacuation Strategy Based on Hidden Markov Risk Assessment Model

地震、海啸、泥石流等自然灾害能够对数据中心产生致命的破坏,数据损坏和丢失会造成巨大的经济损失。对数据中心运营商和用户来说,数据丢失造成的无形损失更加无法承受。因数据中心备份会大大增加数据中心运营成本且影响数据恢复效率,近年来大量研究集中在数据快速疏散方面。遗憾的是大部分研究在缺乏建模和风险量化的前提下假设数据中心的风险为已知因素,存在一定的不合理性。基于SDN网络,本文建立隐马尔可夫模型,动态评估数据中心的风险,根据台风的路径和趋势实时计算台风附近的数据中心节点风险值,把数据实时疏散到风险较小的多个数据中心节点,实现了数据的快速疏散。为了验证模型的性能,本文在Mininet环境下进行了仿真实验,实验结果证明本文提出的模型能够较为高效的完成数据中心风险评估和数据疏散。

Natural disasters such as earthquakes,tsunamis,and mudslides can cause fatal damage to data centers,and data damage and loss will cause huge economic losses.For data center operators and users,the intangible loss caused by data loss is even more unbearable.Because data center backup will greatly increase data center operating costs and affect data recovery efficiency,a lot of research in recent years has focused on rapid data evacuation.Unfortunately,most studies assume that the risk of data centers is a known factor in the absence of modeling and risk quantification,and there is a certain irrationality.Based on the Software Defined Network(SDN),this paper establishes a Hidden Markov Model,and dynamically assesses the risk of the data center to calculate the risk value of the data center near the typhoon in real time according to the path trend of the typhoon.Our proposed strategy evacuates the data to multiple data center nodes with the least risk in real time to realize the data rapid evacuation in a progressive disaster environment.In order to verify the performance of the model,this paper conducted simulation experiments in the Mininet environment.The experimental results prove that the model proposed in this paper can efficiently complete the risk assessment and data evacuation.