Abstract

The Zero-Wait (ZW) policy is widely held to achieve maximum information ‘freshness’, i.e., to achieve minimum Peak Age (PA) and maximum throughput, for real-time Internet-of Things applications. In this paper, it was shown through a series of experiments that the ZW policy is not necessarily the optimum policy for freshness nor throughput in all real-world scenarios. Firstly, the effect of delay on the ZW policy was shown on a LAN. Afterward, the server was located on the Internet, and it was shown that the ZW policy incurred a two-fold PA and throughput performance degradation compared with continuously sending status updates.