Centralized controllers and state estimation strategies are the most common topology implemented for disturbance rejection in smart systems. Although this architecture has been proven to be feasible the advances in computation power and size of logic devices can also enable the implementation of decentralized strategies which, on a large scale, can lead to an overall reliability increase of the system. The goal of this paper is to present this concept within the field of disturbance rejection for smart systems and to explore its capabilities and bottlenecks. A simple example is carried out for a composite plate with two controller/estimator centers. Simulation is used to validate the efficacy of the topology and a design procedure is proposed to guarantee the consensus of the agents within the network.