Product · Alpha

Cairn

Cairn is a self-hosted MQTT 5 broker cluster for IoT fleets. Every device gets its own identity, and you can upgrade a node without the whole fleet reconnecting at once.

Who it is for

Companies with device fleets that want the features of a cloud IoT service on their own Kubernetes or bare metal, and that cannot afford to have the whole fleet reconnect at once every time a broker is updated.

Works with

MQTT 5.0MQTT 3.1.1Kubernetes and HelmPostgreSQLNATSValkeyPrometheus

What it does

Works with the devices you have

MQTT 5 and 3.1.1 over mutual TLS, so devices keep using the MQTT clients they already have.

Upgrades without a reconnect storm

When a node is taken out of service, its devices move to the other nodes at a controlled rate, without a reconnect storm or false last-will messages. This is wired into Kubernetes rolling updates.

Grows beyond one node

Brokers share state through NATS and Valkey. Shared subscriptions and retained messages work across nodes, and each client ID belongs to one node at a time.

A separate identity for every device

Each device gets its own certificate from a private certificate authority, enrolled once with a one-time claim token. Revoking a device also ends its live connection, and a device can only use its own topics.

Device state and rules

Each device has a shadow document for its desired and reported state, in the format users of cloud IoT services know. Rules react to messages and republish them or call signed webhooks, with retries and a dead-letter queue.

An admin console

Overview, devices, certificates with an expiry view, cluster status, rules and a live topic viewer, in English and Turkish, plus Prometheus metrics and runbooks.

How it works

  1. 01

    Enroll

    An admin issues a one-time claim token. The device creates its own key, sends a request, receives its certificate and pins the root CA.

  2. 02

    Connect

    The device connects over mutual TLS. Its client ID has to match its certificate, it can only use its own topics, and the nodes pass messages between each other so devices and services reach each other whichever node they landed on.

  3. 03

    Upgrade

    To upgrade, drain a node. Its devices reconnect to the other nodes, and the node can restart.

Upgrading a broker is the risky moment for a device fleet. When a node restarts, thousands of devices reconnect at once, and the next node can fail under the load. Cairn drains its nodes one at a time so that does not happen, and gives every device an identity that can be revoked.

Cairn is in alpha

Running a device fleet?

Tell us how many devices you have and where the fleet runs, and we will show you what Cairn does today. Pricing and licensing are discussed with each company.

Book a 30-minute demo