A semi-automated reverse engineering method to recommend the best migration-to-cloud strategy

Abstract

in migration of the legacy software to the cloud, the software architecture has to be changed according to the cloud infrastructure and requirements. To achieve this purpose, the software is decomposed (if possible) in to a set of collaborating loosely coupled services to be deployed on the virtual machines. The main rationale behind this decomposition is to provide easy and fast recovery from failed components or replacing the required functionality of the legacy software with the reliable cloud services. In this paper a semi-automated reverse engineering method based on the clustering algorithms is proposed to recommend the best migration-to-cloud strategy. The recommendation is based on four defined metrics: the extent of effort required for reengineering, maintenance costs, achieved availability and the number of cloud services that are used. The proposed method is applied two case studies and the effectiveness of this method is discussed.