As discussed, IoT devices are often resource-constrained; we may not have the luxury of measuring their memory in gigabytes, nor of measuring their processing power by number of cores. Most IoT devices use a microcontroller rather than a full-fledged microprocessor, and run at a few MHz rather than GHz. Specific resource constraints for IoT devices are shown below. Additionally, IoT devices may have physical constraints and accessibility constraints imposed by the operational environment, e.g. pacemakers within the human body. And, finally, cost is a critical constraint for IoT devices. Of course, different devices can be constrained to different degrees – the varied nature of the devices needs to be accounted for in any discussion of IoT. The IETF classifies resource-constrained devices as depicted below.38 Processing Memory Bandwidth IoT Device Resource Constraints Power Storage Figure 4: IoT Device Resource Constraints Classification of Resource-Constrained Devices Class 0 Pre-configured, sensor-like devices that communicate only with gateways and support bareminimum management functionality such as a health indicator or heartbeat. Class 1 Devices that are quite constrained and cannot easily talk to other Internet nodes employing a full protocol stack such as HTTP/TLS but can use protocols designed for constrained nodes and can integrate into an IP network using limited memory, storage, and power. Class 2 Devices that are capable of supporting most of the protocol stacks used on computers. However, even these devices can benefit from lightweight and energy-efficient protocols and from consuming less bandwidth. 38 https://tools.ietf.org/html/rfc7228, IETF RFC 7228, May 2014, Terminology for Constrained-Node Networks 19