In this paper we propose a defense-in-depth IoT architecture that uses multiple layer security measures involving two security mechanisms in discovering and understanding attack vectors. The advantage is that the impact of failure in any one measure is minimized. The defense-in-depth architecture uses firewalls, demilitarized zones, intrusion detection and prevention systems along with associated security policies. For data acquisition and abstraction, we use multiple-tier data models with REST API at the bottom layer and a system process in extracting, processing and feeding data to the application API. Using the newly proposed architecture, we are implementing a water treatment SCADA system that has more than 3,000 PLCs. The data acquisition layer uses US Department of Defense developed API to collect data. The defense-in-depth architecture reduces the risk to IoT networks. Initial tests show that the proposed architecture has the advantages of easy connecting various sensors and reducing the risks of cyber intrusions.