Pervasive Computing envisions distributed applications that optimally leverage the resources present in their ever-changing execution environment. To ease the development of pervasive applications, we have created a pervasive component system (PCOM). PCOM automates the configuration and runtime adaptation of a component-based application using a builtin distributed configuration algorithm. In this paper, we present an architectural extension that allows switching between different algorithms. This enables PCOM to dynamically select an algorithm that suits the computational resources present in an environment. To validate the extended architecture, we compare the overheads of a distributed and a centralized configuration algorithm in two different environments.