From AWS IoT Core

Moving from AWS IoT Core needs the fewest changes: policy JSON imports as it is, shadow topics keep working, and devices keep their AWS-issued certificates.

1. 1Exportread-only, from your AWS account
2. 2Importdry run, then import
3. 3Point devicesnew endpoint and root CA
4. 4Switch backendsSDK import change
5. 5Re-create rulestest each on the bench

#1. Export (read-only)

We give you a small export script. It reads thing types, things, policies, certificates (PEM, status and attachments) and, optionally, shadows using read-only AWS credentials, and writes one JSON file. Nothing in AWS is changed.

#2. Import, dry run first

Upload the file in the console under Migrate, or post it to /v1/migrate/aws/import. The dry run lists what will be created and anything that needs attention. Then import for real.

• Certificates are pinned by fingerprint, so the device certificates work unchanged.
• Policies import as they are; AWS account prefixes in resource names are ignored and policy variables work.

#3. Point the devices

Change the endpoint to <your-endpoint>:8883 and the server root CA to the one from your workspace (or use a server certificate from a public CA your devices already trust). Client ID, key, certificate and topics stay the same. $aws/things/<name>/shadow/... topics are answered by IoT Core twins.

#4. Switch backends

# import boto3
# iot, data = boto3.client("iot"), boto3.client("iot-data")
import tic_sdk
iot, data = tic_sdk.client("iot"), tic_sdk.client("iot-data")

The common thing, certificate, policy, attachment, publish and shadow calls are supported with the same names and arguments. Errors carry a botocore-like response.

#5. Rules

IoT SQL maps closely: SELECT … FROM 'topic/+' WHERE … with topic(n), clientid() and timestamp(). Lambda actions become HTTP actions to your service; SQS and Kinesis become queue actions. Test each rule on the bench before saving.