From Software as a good to software as a service: Preparing the evolution of software products into the cloud

Abstract

Summary form only given. The speed of change in Internet Technologies continues to be impressive and the current “Future Internet” based society increasingly relies on software at all levels. Nevertheless, software systems continue to suffer from aging symptoms caused by the inherent factors of the software modernization process: inevitable; unpredictable; costly; difficult; time- and resource-consuming; poorly supported by tools, techniques or formalisms; poorly understood and maintained; undervalued in traditional software development processes, etc. Furthermore, current technological and business innovations (such as Cloud Computing and Software as a Service (SaaS) concepts) are changing the way in which software is designed, developed, tested, delivered and consumed. All this means that new requirements, never seen before in software development, have to be faced (scalability, multi-tenancy support, usage monitoring, security, SLA compliance, etc.). These challenges are even greater when adapting previously developed software to this new landscape. Different approaches (reverse engineering, ADM, or service-oriented-architecture modernization) have to be analyzed depending on the source language, architecture, target market or available resources. As a result, the complete software life cycle, from requirements to runtime and delivery, has to be re-adapted to the new technological and business conditions, needs and challenges, thus increasing the need of advanced means to support software evolution and adaptation as a key value for the next-generation service-based software modernization. One particular example of complex system modernization is the exposure of legacy systems as SaaS. Model-driven techniques have been used for several years to cope with the challenge of transformation of legacy systems as SaaS. Regardless of the positive features of these techniques, for legacy systems there are many challenges such as are the availability of architectural models describing the system (typically unavailable for legacy systems), the ability to transform the original system into new ones compatible with targeted infrastructures and frameworks, and the support for adaptive deployment. All these features are particular relevant when deploying legacy applications as SaaS on the Web, for instance using different Cloud offerings. Desirable solutions are those that can automate the transformation of existing legacy systems that require being transformed into modern ones exposed as services (SaaS) compatible with modern technology frameworks. This keynote will present the challenges and needs, as well as the results obtained so far in this novel architecture and business model based on the cloud.