Cloud detours: A non-intrusive approach for automatic software adaptation to the cloud

Abstract

A major challenge facing cloud migration is the need to change a legacy (on-premise) application’s source code so that it can better benefit from the inherit cloud computing characteristics, such as resource elasticity and high scalability. When performed manually, those changes are error-prone and may require a great effort from application developers. This paper presents a novel approach to support organizations in automatically adapting their existing software applications to the cloud. The approach is based on the loosely-coupled implementation of non-intrusive code transformations, called cloud detours, which enable the automatic replacement of local services used by an application with similar or functionally-related services available in the cloud. To illustrate the approach, the paper reports on how an initial set of cloud detours, implemented using aspect-oriented programming and a generic cloud library, was used to seamlessly adapt an existing file-based Java application to save application data in a cloud-based storage service.