Optimality of distributed state estimators with communication cost : A case study