Most wireless sensor networks are used for monitoring purposes, where some nodes sense the environment and forward their measurements to an entity called the sink. Nodes around the sink tend to have their energy depleted faster than other nodes in the network, as they have the burden of forwarding the measurements from all the nodes. Thus, the current trend in wireless sensor networks is to consider networks with several sinks. In this context, anycast communications are important: each node producing traffic has to send its measurements to any available sink, generally to the closer one. In this chapter, we aim at reducing the congestion in the whole network by jointly selecting sinks and routes from sources to sinks. We show that considering the two problems separately results into low performance. After proving that the joint sink and route selection problem is NP-hard, we propose a mixed integer linear program in order to obtain optimal solutions. Then, we propose a centralized mechanism that selects sinks and routes packets with reasonable performance (compared to the optimal solutions) and control overhead. Our approach contributes to balance the generated traffic in the network by considering special nodes called pivots. Our mechanism is validated by extensive simulations on generic network topologies. Then, we show how this mechanism can be distributed, using only a limited knowledge.