Crate fakecloud_iot 

AWS IoT Core (iot) restJson1 control-plane service for fakecloud.

The full 272-operation AWS IoT Core Smithy model (SDK id IoT, SigV4 signing name iot, endpoint prefix iot). Every operation is a RESTful <METHOD> <@http URI> route with path labels (e.g. POST /things/{thingName}, GET /policies/{policyName}, PUT /jobs/{jobId}, POST /rules/{ruleName}) so requests are routed by their HTTP method + @http URI template. The route table, per-operation HTTP bindings, model-derived input constraints, and output member shapes are all generated from the Smithy model (see src/generated.rs, produced by scripts/generate-iot-tables.py), so the registry, jobs, rules, and security control plane tracks the model exactly.

What is real. Things, thing types, thing groups (static + dynamic), billing groups, policies (+ versions + attachments), certificates (+ CA certificates + principal attachments), jobs (+ job templates), topic rules (+ rule destinations), Device Defender security profiles / scheduled audits / audit configuration / mitigation actions / custom metrics / dimensions, provisioning templates (+ versions), domain configurations, fleet metrics, role aliases, authorizers, streams, OTA updates, packages (+ versions), certificate providers, commands, and every other modelled resource mint proper ARNs / ids, persist their attributes, echo them back on read / list with round-tripping pagination tokens, and enforce referential rules (attach / detach principals + policies, thing-group + billing-group membership). State is account-partitioned and persisted across restarts. CreateKeysAndCertificate / CreateCertificateFromCsr mint a 64-hex certificate id + ARN and a structurally-shaped PEM certificate and RSA key pair. DescribeEndpoint returns a deterministic account-specific endpoint for every endpoint type (iot:Data-ATS, iot:CredentialProvider, iot:Jobs, …). Jobs carry a lifecycle (QUEUED -> IN_PROGRESS -> COMPLETED); topic rules store their SQL + actions verbatim. Input validation is model-derived: required members, string @length, numeric @range, and @enum constraints are enforced with IoT’s declared exceptions (ResourceNotFoundException, InvalidRequestException, ResourceAlreadyExistsException, DeleteConflictException, VersionConflictException, …).

Honest emulation choices (documented, not stubbed):

• There is no live MQTT broker or device connectivity. The registry / jobs / rules / security control plane is fully real and persisted, but no message is routed, no topic-rule action is executed against a real target (SNS / SQS / Lambda / …), and no device ever attaches over MQTT. The data plane lives in the separate fakecloud-iotdata service (device shadows + retained messages).
• SearchIndex and the aggregation queries (GetCardinality, GetPercentiles, GetStatistics, GetBucketsAggregation) run against the in-memory thing registry with a bounded query subset (a thingName: prefix / exact match and the wildcard *). Queries outside that subset return InvalidQueryException rather than a wrong result.
• Certificates are structurally-valid PEM placeholders, not real CA-signed X.509 chains; the key pair is a shaped placeholder, not a usable private key.

Re-exports

pub use service::IotService;

pub use service::IOT_ACTIONS;

pub use state::IotData;

pub use state::IotSnapshot;

pub use state::SharedIotState;

pub use state::IOT_SNAPSHOT_SCHEMA_VERSION;

Modules

generated

persistence

Snapshot save/load for AWS IoT Core control-plane state.

service

AWS IoT Core (iot) restJson1 dispatch.

state

Account-partitioned, serializable state for the AWS IoT Core control plane.

validate

Model-derived input validation for AWS IoT Core operations.