Peer-to-Peer (P2P) backup systems have become popular solutions to ensure data availability. The key idea of these systems is to use the shared memory space between peers to store data. Nevertheless, some limits appear because of the large scale and the dynamic context of P2P networks and the lack of collaboration between peers. Likewise, as development data become more volumetric and distributed on multiple sites, these data need an optimized way of backup. Within the framework of concurrent engineering, our motivation is to create a corporate environment that facilitates communication and collaboration between peers to better support large scale, highly dynamic systems and handle the backup process in an efficient and organized way. In this paper, we propose a collaborative-based approach allowing efficiently exploiting resources of each peer. Our solution is based on a specific clustered organization of the network in order to improve peer interactions. Performances are validated throught simulations.