基于数据流的软件抗衰重启技术研究

制定准确实时的细粒度软件抗衰策略可以进一步增强策略的准确性,优化重启效果,提高软件可靠性。在数据流的基础上,结合了软件抗衰重启理论,定义了基于数据流得软件系统模型,制定了较为全面的系统性能参数,并且通过异常模式提取和趋势分析的方法有效地对软件异常子系统进行预测,确定出系统的重启层;同时给出了新型告准确度的重启相关度的计算方法,增强了其实时性和准确性,从而为执行高效准确的细粒度软件抗衰提供支持。

Research on Software Rejuvenation Based on Data Streams

In order to advance software availability and reliability, improve the effect of the software rejuvenation and boost up the veracity of the policy, the real-time fine-grained software rejuvenation policy would be established. Based on the data streams, the theory of software rejuvenation was integrated, the model of software system was defined based on data streams, and the performance features of system was put forward. By extraction and trend detection of unusual patterns based on the data streams, the sub-system was predicted which should be executed with software rejuvenation, the restart levels were decided, and a new method was set down to calculate the degree of restart dependence with high veracity, which enhanced the characteristics of real time and veracity and supported to execute fine-grained software rejuvenation policy with high efficiency.