A cloud platform architecture recovery metric method

Cloud platforms have become the most popular architecture for providing software services today, and traditional architecture evolution evaluation methods cannot achieve better results on cloud platforms because they are different from the architecture of traditional software. In this paper, we propose a method for architecture recovery based on the decentralized characteristics of cloud platforms. First, the method identifies the inapplicability of traditional monolithic software architecture recovery methods under the cloud platform architecture, and then a fast recovery method for cloud platform architecture based on the new city of dispersion features. The process of cloud platform architecture recovery extracts the overall architecture information, integrates the transformed New Town association dependency graph based on the architecture information, and then uses the architecture dependency graph as the input for distance-based aggregation of decentralized components. Finally, the above method is applied to three cloud platform projects, and the evaluation of the effectiveness of architecture recovery is verified separately. The experimental results show that the cloud platform architecture recovery method based on decentralized features can fully extract the cloud platform architecture information and achieve the accurate metric of architecture description.