In Wireless Sensor Networks (WSN), a tradeoff between the network lifetime (limited by energy of sensors) and redundancy of actively sensing and communicating sensors (implicated by coverage requirements for the measured area) has to be established, typically in an over-deployed environment. This is achieved by scheduling algorithms which periodically alternate the state of sensors between ”asleep” and ”awake”. Obviously, the period length of the periodical synchronized scheduling affects the network performance and lifetime. Controlled Greedy Sleep algorithm is a quasi-optimal synchronized sensor scheduling algorithm which increases network lifetime while maintaining correct functionality, based on local decisions of sensors. This paper investigates the optimization of the period length of this algorithm and highlights best practices with simulations. Studies have been performed within a ring topology and in random square topology.