Constructing and Generating a Time-varying State- Space Vector against Cloud Service Event Interactions
Yiyi Jiang, Xiaojian Li, Chunhe Xia, Hangping Hu, Zhen Zhang, Hailan Wang · 2021
the abnormal interaction between cloud service events is an essential factor leading to failure. Using neural networks explores the interaction between cloud service events, which needs a method to transform the cloud service events into a vector. However, cloud service events exist as a log with time- varying characteristics. The dataset of cloud service is vast, including numerous strings that cannot be directly involved in calculations. There is a lack of pathway transform data semantics to interaction semantics. Thus, this paper provides a method to construct cloud service logs into vector. Regarding cloud service events as natural language, the method characterizes cloud service events as vector without building a corpus. In addition, we use the semi-supervised neural networks to predict the interaction between cloud service events. The results show that the effectiveness of our method, indicating that it meets the needs of project. This work also provides a foundation for security traceability.