Keynote: Intelligent Internet of Things: Start Analyzing Your Global Device Data for Real-Time
summary
Claims a platform that connects millions of devices bi-directionally using Cloud IoT Core, Pub/Sub, BigQuery and Dataflow and runs ML with Cloud ML Engine and TPUs; states IoT Core went GA in February and handles MQTT/TLS with JWT-signed tokens so tiny 8‑bit MCU + Wi‑Fi + crypto can authenticate with a ~10 KB stack. Gives data points: paper machine 4 TB/year, Dexcom ~10 TB/month uncompressed, cars 22–25 GB/hour, toy line 4 GB/day. Shows smart‑parking cut deployment from 1–2 weeks to 2–3 days and a molding line saved 3,500 hours (~$1M).
AI-written summary · about
chapters (9)
- 0:00Intro
- 1:08Use Cases
- 7:31Master Big Data
- 11:56Security
- 14:14Insights
- 17:09AI
- 19:07Tensor Processing Units
- 20:13Odin Technology
- 22:07Conclusion
related talks
52mMachine Learning in IoT Applications
21mKeynote: Future of IoT: A VC Perspective
28mOpening Keynote: TCO of IoT, new managed gateway and more
20mFrom Cloud to Fog Computing and IoT
48m50 Billion Connected Things are Coming
21mKeynote: Can the Web Save the Internet of Things?
47mLessons and Advances in IoT: Toward 1M Connected Devices
27mThe IoT storm... Will we survive the coming deluge?