We consider two models for estimating the expected states of nodes in networks where the observations at nodes are given by random states and measurement errors. In the first model, we assume independent successive observations at the nodes and the design question is how often the nodes should...