Skip to content
Greengrass V2 Edge Walkthrough

Greengrass V2 Edge Walkthrough

A greenhouse edge lab in two acts — closed loop on the LAN, then wire the same data into AWS. Every step is plain AWS CLI.
Get startedJump to AWS wiring

What you build

Section titled “What you build”

One NUC
Intel NUC6CAY — compact x86 PC used as the Greengrass core in this lab (spare unit on hand; any quiet Ubuntu 24.04 x86 host works). runs Nucleus
Greengrass Nucleus — the edge runtime on the core device. Installs components, talks to AWS, and hosts the local client-device stack.. Up to three ESP32-S3
Espressif ESP32-S3 — Wi‑Fi microcontroller used here as the Greengrass client device (not a second core). boards are zone client devices
Greengrass client device — an IoT Thing (here the ESP32-S3) that discovers a core and uses local MQTT. It does not run Nucleus.. Press BOOT → green RGB on that zone. Chip temperature streams every 10 s. AWS archives it, tracks state, alarms on heat, commands the RGB blue from the cloud, and trains a model the core runs locally (red RGB).

No Terraform. No provision wrappers. Visible AWS CLI
AWS Command Line Interface v2 — every cloud and deployment step in this walkthrough is a visible aws command, not a wrapper script. and JSON under artifacts/.

RoleHardwareSoftware
Core
Greengrass core device — the Linux host running Nucleus (this lab: Intel NUC6CAY). Client devices discover and connect to it.

NUC6CAY
Intel NUC6CAY — compact x86 PC used as the Greengrass core in this lab (spare unit on hand; any quiet Ubuntu 24.04 x86 host works)., Ubuntu 24.04

Nucleus, Moquette, bridge, GgEdgeLoop, ZoneAnomaly

Client
Greengrass client device — an IoT Thing (here the ESP32-S3) that discovers a core and uses local MQTT. It does not run Nucleus.		ESP32-S3 DevKitC-1 (up to 3)Firmware — button, chip temp, RGB
ESP32-S3 zonesBOOT · temp · RGB
→
NUC · Greengrass V2
MoquetteMQTT BridgeGgEdgeLoopZoneAnomalyLogManager
→
IoT Coretopics
→
S3telemetry lake
DynamoDBzone state
CloudWatchmetric · alarm · logs
Control · blue RGB
Alarm
→
SNS
→
Lambda
→
IoT Core
→
GgEdgeLoopblue RGB
ML · red RGB
S3 lake
→
SageMaker
→
S3 artifacts
→
ZoneAnomalyred RGB
Zones stay on the LAN. The bridge is the only path into IoT Core. Rules fan out telemetry; alarm and SageMaker close the loop back to the zones.

Path

Section titled “Path”

Read Concepts and Why Greengrass? once, then run the numbered pages in order.

Before you begin

Section titled “Before you begin”
  1. Tooling — AWS CLI, jq, ESP-IDF
  2. Prerequisites — account, BOM, naming
  3. Concepts · Why Greengrass?

1 · Greenhouse on the LAN

Section titled “1 · Greenhouse on the LAN”

Local sense → actuate. The cloud may mirror topics; nothing consumes them yet.

#PageYou leave with
1IdentitiesThings, certs, policies, token-exchange role
2Prepare the coreUbuntu on the NUC, walkthrough.env
3Install NucleusHealthy core device
4Local MQTT stackMoquette, Auth, Bridge, IPDetector
5Associate zonesDiscovery returns the core host:8883
6Loop component

GgEdgeLoop RUNNING on the NUC

7Flash zonesBoards on Moquette; telemetry every 10 s
8Prove the LAN loopBOOT → green RGB on that zone only

2 · Wire into AWS

Section titled “2 · Wire into AWS”

Same topics. Now AWS consumes them, and the cloud can drive the zones. Full topic-by-topic wiring: Architecture.

#PageProof on the bench
9Deploy from the cloudComponent from S3; logs in CloudWatch
10Archive telemetryJSON objects under telemetry/ in S3
11Zone stateLatest temp + who drove the RGB in DynamoDB
12Raise an alarmThumb on the chip → email
13Cloud commandsAlarm turns that zone’s RGB blue
14Edge inferenceCore turns RGB red before the cloud alarm
15TeardownEvery resource gone

RGB tells you which layer acted: green = button loop, blue = cloud command, red = edge model.