Remote-controlling Hardware with Intel ME / AMT
Our Test Laptop (ThinkPad X240), as well as most other Intel Systems produced since the early 2010s, come with the so-called Intel Management Engine, which implements Intels Advanced Management Technology. Essentially, this is a piece of proprietary firmware living on the network interface card, which is able to monitor and remote-control most parts of the system over a network connection. Since the ME runs separately from the CPU, it is active even when e.g. the system is powered down. For us, the ME could be quite useful for in-hardware debugging. For example, you can use it to tunnel a serial COM port over Ethernet, which is really useful for systems which lack an external serial connector.
Preparation
First, you need to enable the ME in the targets BIOS / firmware settings. After that, you can enter a separate ME setup screen, in which you have to configure an access password, and a network connection (usually a static IPv4 address). Currently, we use static IPv4 addresses in the 10.0.0.0/24 subnet. It turns out that the ME has its own identity on the network, separate from the main system. This means that it can have a separate IP address or even subnet from the rest of the computer. Confusingly, it can also have the same IP address as the main system, with the same MAC address, but still behave as a separate network node.
To be able to connect to the target system from your developer system, you need to configure your own network interface to lie in the same subnet. When using static IP addresses, you might want to disable your NetworkManager/DHCPCD on the interface.
Direct Access
When the ME is enabled, the target system provides a web interface on a number of ports. The most common one you will need is Port 16992.
Access via a unified web application
Install a Node.js package called mesh-commander on your development machine and execute it, to open a nicer web application on your local port 3000.