Abstract

this paper, we posit that an explicit architectural focus can remedy many of these difficulties and enable flexible construction and evolution of large systems. Software architectures present a high level view of a system, enabling developers to abstract away the irrelevant details and focus on the "big picture. " Another key is their explicit treatment of software connectors, which separate communication issues from computation in a system. However, existing architecture research has thus far largely failed to take advantage of this potential for adaptability, for two reasons: . connectors are often not treated explicitly or, when they are, they are too rigid and do not accommodate modification of their attached components easily; and . no specific techniques have been developed to support flexible architecture-based design and evolution. Three distinct building blocks of a software architecture are components, connectors, and architectural configurations (topologies) [13]. Each of them may evolve. Our work to date has focused on the evolution of individual components and architectural configurations. In the future, we intend to investigate the proper techniques for evolving connectors. For evolving individual components, our approach expands the traditional techniques for supporting evolution (e.g., modularity, typing). We introduce explicit, flexible connectors to aid the evolution of architectural configurations. The following section discusses our approach to component evolution and introduces an architectural type theory on which the approach is based. Section 3 discusses the role of software connectors in the evolution of architectural configurations. Conclusions and a discussion of ongoing work round out the paper.

Citations