Recently, many open source software have been used under several situation. Especially, several open source software are implemented under the network computing areas of edge computing and clouds. We propose the data preprocessing and fault big data analysis based on deep learning for open source software. In particular, we analyze the open source fault big data from various points of view. Also, the fault data sets are analyzed by using the frequency encoding and count encoding from the special characteristics of fault detection phenomena for open source software. Moreover, we discuss the characteristics of fault big data in open source software. Furthermore, several numerical examples for fault big data analysis in open source software are shown in this paper. Then, we focus on several open source software. We compare the difference of the deep learning and the stochastic models from the standpoint of fault estimation.