Enabling Optional Software Components In Multi-Constraint Deployment Optimization Problems

This paper presents methods for adding optional software components to a Distributed Real-Time Embedded (DRE) deployment optimization problem, and methods for attempting to enable deployment of the maximal number of optional software components. By building upon prior work that allows rapid deployment optimization, one can predict probability of success with high confidence, and utilize that information to generate a lower bound on number of times deployment of an optional component should be attempted. This paper also introduces three algorithms for selecting which optional components should be attempted in a deployment, and experimental data which shows that the order in which optional components are considered has a high impact on the maximal number of optional components that can be included in a deployment. Keywords-deployment, optional software components, distributed real-time embedded systems