Urgent.News

What's breaking now, across thousands of outlets.

Tech

CoAP: 2,188,380 answers for a protocol designed to be spoken to by anything

CoAP: 2,188,380 answers for a protocol designed to be spoken to by anything A ZoomEye query for service="coap" returned 2,188,380 matches on 28 October 2026. CoAP is the Constrained Application Protocol, standardised as RFC 7252 in 2014, and it was built for devices that cannot afford TLS handshakes or verbose headers: sensors, meters, building controllers, industrial gateways. Every one of those…

CoAP, also known as the Constrained Application Protocol, is a standardised protocol for devices such as sensors, meters, and building controllers. It is designed for devices that cannot handle TLS handshakes or verbose headers. CoAP uses UDP port 5683 and does not require a connection handshake or a banner, distinguishing it from port-only queries.

The protocol uses multicast capability, allowing a single request to a group to be answered by every device in it. However, CoAP has no built-in authentication, relying on optional security measures like OSCORE and DTLS. The discovery resource, defined as /.well-known/core, is used for device enumeration and can reveal the full interface list of a device without requiring any credentials.

This makes it a potential security risk as it allows an arbitrary source to access the full interface list of a device. The 2,188,380 figure represents the number of devices that answered a CoAP protocol probe, indicating that a large number of constrained devices are reachable at the network layer by something other than their intended client.

To mitigate this risk, query your own address space for devices using CoAP and check their /.well-known/core resources to understand what the device exposes. If the device only speaks plain CoAP on a routable network, it lacks confidentiality for its commands. Finally, identify the vendor and firmware from the CoAP response options and look for vendor guidance on OSCORE support and disabling multicast responses.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

Coverage Is Not Capability: What Whisper Taught Me About African Speech

A language appearing in a model’s supported list says little about whether the model will work for your speakers. Why Whisper struggled with the speech I needed, and what I now check before building…

  • Whisper trained on 680,000 hours of audio, but African languages underrepresented
  • Swahili has ~5.4 hours of data, Dholuo not listed, highlighting coverage disparity
  • Evaluation should consider real-world audio, licensing, and deployment environment

More from Tuesday 27 October →