Do you deploy 10’s to 1000’s of devices and need to quickly identify them in the field for configuration, maintenance, or troubleshooting? This article explores how this can be made much easier by using device IDs that are easy for humans to process.
In a previous article (A Survey of device ID mechanisms), we explored some of the commonly used options (MAC addresses, UUID, 2D barcodes, etc.). However, most of these are optimized for consumer devices and scenarios where machines process the IDs.
Why are human-friendly IDs important? Consider a few scenarios:
- 8 identical temperature sensors are located at various places on a piece of equipment, and we need to identify the location of each one.
- 3 IO modules are attached to a gateway, and we need to select which module a sensor is connected to when creating a virtual component in a dashboard.
- One of the above 3 IO modules in a crowded control panel has failed, and we need to replace it. How do we identify it?
- A user calls with a technical support problem on a piece of equipment, and we want to quickly look up the details on the hardware they have (model, version, etc).
- We need to list and verify the device serial numbers on a packing list for a shipment to a customer.
- Devices are provisioned and configured in the shop for several sites, and during installation, we need to make sure we install the right device in the right location.
- A device’s mDNS name defaults to the device’s serial number. Once a device is attached to a network, the user needs to access the Web UI using the address:
http://<device id>.local. It is helpful if the ID is simple enough for a user to read from a label and easily type into a browser.
In the real world, especially the industrial world, there are many cases where humans need to interact with equipment beyond the initial provisioning.
The problem with (only) QR Codes
QR Codes are increasingly being used to identify devices. We buy a new device, scan the QR code with an app on our phone, and the device gets added to the system — pretty slick! QR Codes are optimized for provisioning, with data flowing in one direction — from the QR code to the system. However, using the QR code to identify a device that is already deployed in the field is more challenging. We need to be able to look at multiple devices in the field and quickly identify them. Also, consider the scenario where you are on the phone supporting a customer — large QR codes are not transmitted over the phone very well. QR codes alone do not solve every problem.
What works well in the consumer space is not always appropriate for the industrial space. While there are many consumer technologies we can leverage, what Google, Apple, and Amazon are doing is not automatically the best solution for industrial applications. It may be possible to identify a device using a QR code using a complex standard like Matter, but do we have the development resources to make this happen? Do we need the complex security and privacy features in an industrial context? Is there a simpler/better solution?
What makes an ID easy for humans to process?
Many MFA (multi-factor authentication), one-time passwords, and authenticator apps use codes that are composed of 6 digits split into two chunks:
From experience, we know these numbers are easy to remember and type into a form. Even on our phones, it is often more work to copy and paste than simply look and remember. Why is this? Human short-term memory can typically hold about 7 items at once (plus or minus 2). Chunking these numbers further reduces cognitive load for users. For example, “283-178” is easier to remember than “283178”.
Human-friendly IDs (HFID)
If we start with the format used by authorization codes (XXX-XXX), which is six digits, this gives us one million possibilities (10^6). That is a lot of devices, and may be adequate for some applications.
If we expand this ID to also include capital letters (example: N8H-9MJ), then possible combinations are over 2 billion (36^6).
If we eliminate letters and numbers that are easily confused (0, O, 1, I, 2, and Z), then reducing the base from 36 to 30 characters (30^6) gives us ~729 million combinations. Randomizing the characters in the ID also helps ensure that the IDs are visually unique and not easily confused with IDs on neighboring devices. We might also want to require a letter and a number in each 3-character chunk to make the ID easily recognizable as an HFID, which would reduce the number of combinations to ~340 million.
340 million is far more than most organizations in the industrial space will ever produce. If there is ever a need for more IDs, then the number of characters could be expanded with a minor loss in usability (example: PU89-V6M2), which gives us 656 billion combinations (30^8).
Managing Device IDs (hfid.org)
This scheme does require a central authority to create these IDs — ideally, an API that various manufacturing processes (including subcontractors) can interface with and pull new IDs in real-time. The real value occurs when multiple companies can use the same ID registry, and thus devices from different manufacturers can interoperate. For this, it would be beneficial if there were a vendor-neutral organization that generated these IDs on demand. We created a prototype of a system to do this.
Once your account is approved, you can generate new HFIDs via the website or API. Check it out …
Why not both?
Why not both an HFID and a small QR code containing the HFID on your device? Then it is easy for both humans and machines to process the ID.
Summary
Human-friendly device IDs are a way to uniquely identify a large number of devices, yet make it easy for humans to quickly and accurately identify devices in the field. Similar to Internal Part Numbers designed for humans, systems designed to work well with humans are key to reducing mistakes and improving efficiency.
Even if you are using devices/sensors/equipment manufactured by others that already have an ID, adding an HFID to any device you deploy allows you to identify it with less effort.
The human eye and brain are remarkable instruments. Implementing IDs that are easy for humans to process has countless benefits as you manufacture, track, deploy, maintain, and service your products, devices, and equipment. Quickly identifying a device is the first step to doing anything with it.
Don’t just use an ID, use an HFID.