Defining Dependable Dynamic Data-Driven Software Architectures

The thesis of this vision paper is that the Dynamic Data Driven Applications Systems (DDDAS) is a promising paradigm to adopt for assisting architectures to self-maintain their dependability properties, as the software architecture tends to evolve in response to changes in the operating environment, changes in contexts, and dynamic usages of the application. In this perspective, the architecture becomes an integrated computational and measurement artifact aimed at measuring, simulating, and controlling the runtime evolution of dependable software systems. This perspective is novel and has the promise to form a built-in support for the runtime dependability analyses, reasoning, and evaluation for many architecture-centric approaches such as product-line, service oriented, and model-driven paradigms. The contribution of this position paper is a definition of Dependable Dynamic Data-Driven Software Architectures (DSA), inspired by this paradigm. We describe the major components which can "orchestrate" to realize DSA. We highlight some challenges and opportunities.