Ad-Opera: Music-inspired Self-adaptive Systems

摘要

These days, systems are emerging as agglomerations of software, hardware and people. They are highly distributed, heterogeneous, context-aware, mobile, and adaptive to resource availability and requirements evolution. New computing paradigms for such systems are required. Ad-Opera proposes radically new approaches to software engineering, based on investigation of, and inspiration derived from, a specific class of self-adaptive systems: adaptive music. In adaptive music, the interaction with the audience and with the environment plays an active role in the composition process. The musical work is fully defined only at execution time depending on context available resources, and only after appropriate user intervention. Nevertheless, the self-adaptation must ensure the character of the music desired by the artist, called style. The Ad-Opera new computing paradigm will yield the following main novelties: (i) late specification that permits the under-specification of parts of the system by delaying their completion to runtime, (ii) context as first class entity that tightly couples the context to the system computational state, (iii) explicit support for self-adaptation that views adaptation as a pervasive normality in the system's behavior, rather than as the exception.