Beyond Simulation: Composing Scalability, Elasticity, and Efficiency Analyses from Preexisting Analysis Results

Abstract

In cloud computing, typical requirements of Software-as-a-Service (SaaS) applications target scalability, elasticity, and efficiency. To analyze such properties, software engineers need efficient specifications that acknowledge for uncertainties within the underlying cloud computing environment. However, existing analysis specifications are based on simulating the system as a whole, which is inefficient and requires full knowledge of the underlying environment. To cope with this problem, we envision to structure systems in independent operations, each annotated with novel scalability, elasticity, and efficiency attributes from preexisting analyses, e.g., conducted by engineers that had sufficient knowledge of the environment. Such attributes enable highly efficient compositional analyses of the system as a whole. In this vision paper, we describe our initial ideas for our new composition approach based on a simple running example.