Creating an IoT Data Layer for Collecting, Storing, Analyzing
summary
Shows a working Linux-based edge that runs the same Influx stack edge-to-cloud (InfluxDB, Telegraf, Kapacitor, Chronograf), collects time-series from many protocols and forwards five-minute rolling averages to the cloud while keeping high-frequency data local for immediate reaction. Demonstrates CO2 safety automation with local Kapacitor alerts and a Pine64 edge box that runs a year of 1 Hz data from ~24 devices × 5 streams using 64 GB SD at ~20% capacity and ~170 MB stored in the cloud. Telegraf ingestion and Kapacitor alerts are used for sync and local reactions.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 0:50Data Layer for IoT
- 3:34Collecting in the Cloud
- 5:23Collecting at the Edge
- 7:58Distributed Data Collection
- 10:42Data Layer Architecture
- 11:51Data Management
- 13:47IoT Architecture
- 18:05IoT Dashboard
- 19:25Time Series Data
- 21:18MongoDB
- 26:15Questions
- 34:45Two redundant topology
- 36:40Remove duplicate data
- 37:40Synchronize data
- 39:15Telegraph
- 40:15Influx
related talks
45mIoT and Stream Processing
16mEdge to Cloud with InfluxDB
06mUnderstand your Data with InfluxDB Cloud
07mLightning Talk: Data Visualizations with InfluxDB: Integrating plotly.js
27mStream, Store, Visualize: Fast-Track Real-Time Analytics the OSS Way
35mTake IoT to a New Level with
45mA Practical Guide to Dashboarding
45mUsing time-series data to better understand our products & infrastructure