Cloud edge cooperative attack identification method and system, computer equipment and storage medium

The cloud edge cooperative attack identification method comprises the following steps: an edge end trains an attack identification model based on a GNN algorithm and transmits self training set data to a cloud end; the cloud updates a self data set based on the training set data and trains an attack recognition model by using the data set; the cloud side screens the attack recognition model and distributes the screened model to the edge side; in the whole process, the edge end collects hostile attack traffic records to update a self training set. According to the method, GNN and an open source data set are utilized to train an attack recognition model, edge computing ends distributed with certain resource training capability and information collection capability are trained, the edge ends with high recognition precision transmit the training set to the cloud to update the training set of the cloud, and the cloud issues the model with high recognition precision to the cloud, so that the recognition efficiency