Software Producing Organizations (SPOs), i.e., open source and commercial organizations producing mass-market software, are increasingly engineering Very Large Software Systems (VLSSs). These VLSSs are, however, directly conflicting with current trends in software engineering: development and deployment agility, ecosystemification, inter-organizational collaboration, and multi-tenancy. These trends could be highly beneficial for VLSSs, but few to none techniques exist to exploit these trends. We see these trends as a call to action for software engineering researchers to rise up to the challenge of developing a framework that embraces the needs for variability, openness to collaborative software development and composition, and resource sharing across highly specialized applications. We envision the AMUSE framework as the answer to this challenge, a framework that enables VLSSs in the software engineering context of the future.
The AMUSE framework specifically addresses the (1) composition, (2) configuration, (3) deployment, and (4) monitoring challenges of VLSSs will be faced by (1) designing and developing a composition language for software composers, by (2 & 3) developing shortest-path algorithms for configuration upgrades of highly dependent and integrated systems, and (4) by providing a set of software operation knowledge tools that observe and analyze behavior, through process mining, of MDD software systems in the cloud. The ultimate goal of such a framework is to provide a self-healing system that, based on predefined policies, can change its own deployment in the heterogeneous cloud.
More details about the AMUSE project can be found here PDF