IoT on Google Cloud Platform
summary
Shows a working pipeline: Zigbee mesh → Gateway using MQTT clients → RabbitMQ on Compute Engine with an MQTT→Pub/Sub adapter, then App Engine subscribers for processing. Load tests run with simulated gateways (template VM = 100 gateways per VM, 4 messages/sec each) to measure throughput, mobile data usage and cost. Uses Cloud Storage, Datastore, Pub/Sub and Stackdriver for monitoring, relies on instance groups for auto‑scaling and RabbitMQ disc nodes for durability; architecture proves scalable though Pub/Sub lacks native MQTT.
AI-written summary · about
chapters (22)
- 0:00Intro
- 1:05About augusta
- 1:38augusta development team
- 2:11IoT projects
- 2:57Opportunities
- 4:00The Reality
- 4:33Whats Google up to
- 7:06How we got into IoT
- 10:57Google Cloud Platform
- 13:55Challenges
- 15:38Scenario
- 16:21Parking Ramp
- 17:51Street Lighting
- 18:07Network Protocol
- 18:49MQTT
- 19:37Message Broker
- 24:24Queueing
- 25:48MQTT Topics
- 27:55Gateway Simulation
- 32:05RabbitMQ
- 38:16Compute Engine Dashboard
- 43:05Load Testers
related talks
36mHow to Build a Scalable Serverless IoT Architecture on GCP
45mIoT at Google Cloud
33mManaging IoT Storage with Google's Cloud Platform
40mBuilding Scalable Resilient IoT Messaging
40mIoT with Sigfox and Google Cloud
25mAn Introduction To Google Cloud Platform
20mStreaming Processing Options in Google Cloud
57mDebugging Java in Production on Google Cloud Platform