Deployment Time Optimization of Distributed Applications

This paper proposes a novel optimization technology called deployment time optimization and presents several examples where the proposed technology can improve the performance of distributed applications. Using the configuration information collected from the target operation environment of an application, the proposed optimization approach tries to deploy only necessary components and best alternatives for the target environment. Since deployment time optimization works at module level, we may call this approach a macro level compilation. To enable deployment time optimization, we have designed a novel framework called Blue Pencil. The Blue Pencil framework consists of policy module for storing rules for optimization, programming environment to support generating glues between distributed modules, configuration discovery module that can collect information about a target environment, and code transformer that can customize an application for a target operation environment. We conclude this paper by presenting a simple example on how to select appropriate bindings between distributed objects as a proof of concept.