Energy efficiency is the main concern of applications supported by WSNs where some nodes are battery operated and it is difficult, expensive or even impossible to renew their energy. In the literature, four classes of energy-efficient techniques have been distinguished in WSNs: (1) topology control where the transmitter adapts its transmission power, depending on the receiver (2) energy-efficient routing where the energy consumed by an end-to-end transmission is minimized and nodes with low residual energy are avoided, (3) node activity scheduling where nodes come back to the sleep mode to save energy and finally (4) optimization of transfers in order to avoid useless transfers. The OCARI WSN has been designed for industrial applications where energy efficiency matters. OCARI targets applications such as monitoring of industrial equipments or civil engineering, performance testing of equipments, radioprotection of site maintenance and state control of some devices. Such applications require determinism in the medium access and low energy consumption. OCARI is based on the 802.15.4 physical layer and supports the application objects defined by ZigBee. It differs from other protocols like ZigBee, WirelessHART and ISA100.11a by its unique characteristics: - An energy efficient deterministic medium access supporting relaying of time-constrained packets, called MaCARI. - A proactive routing taking into account the residual energy of nodes, minimizing energy consumption and supporting nomadism, called EOLSR. The residual energy of a node is estimated by a model that provides an accuracy better than the linear model, especially in the case of a low duty cycle. - A node activity scheduling based on node coloring, reducing interferences and thus optimizing energy efficiency, called SERENA. The OCARI topology consists in a mesh of stars, each star is ruled by a coordinator node. Coordinator nodes are able to route packets whereas simple sensor nodes are not, they are only able to communicate with their coordinators. In OCARI, both types of nodes can sleep, while maintaining network connectivity. In conclusion, OCARI optimizes transfers with collision avoidance, reduction of interferences, and broadcasts optimization. Cross-layering is used to reduce the routing overhead by maintaining only useful routes. Moreover, it allows OCARI to optimize data gathering delays and data freshness.