Apache IoTDB vs ThingSpeak

Apache IoTDB is an open-source industrial IoT time-series database for collecting, storing, managing, and analyzing device data across edge and cloud deployments. The project describes high-speed writes for millions of networked devices alongside fast reads. Query features include aligning time across devices and sensors, time-series computation, and aggregation by time dimensions. Ingestion options include IoTDB-SQL, JDBC, Java, C++, Python, and Go native APIs, REST, MQTT, CSV, and TsFile. Named analysis ecosystem examples include Hadoop, Spark, Flink, and Grafana. Deployment options include cloud installation, a desktop terminal tool, and synchronization between cloud and on-premise machines. The REST service is disabled by default and uses Basic Authentication except for the health check endpoint. Data and cluster permissions can be managed through users, roles, and privileges. The homepage states disk storage can cost less than $0.23 per GB. The project lists energy, aerospace, transportation, steel, and other IoT scenarios. Its JDBC documentation cautions that insert operations may not provide high-performance writing and recommends the Java Native API for Java applications.

ThingSpeak is a cloud IoT analytics platform for collecting, visualizing and analyzing live data streams. Devices can send data through REST or MQTT APIs, and listed libraries support Arduino, ESP8266, ESP32, Particle and Raspberry Pi. Each channel stores up to eight fields, with data private by default and an option to make channels public. Users can run scheduled MATLAB code hosted by ThingSpeak without a separate MATLAB license. React and TimeControl apps support responses to channel conditions and scheduled computations. The platform supports near-real-time stream processing, custom dashboards and machine learning tools. MathWorks describes its uses as IoT prototyping, proof-of-concept systems and hosted production for small- to medium-sized systems. The Free plan is free for non-commercial use, with four channels, three million messages per year and a 15-second update interval. MATLAB code execution is limited to 20 seconds for free users; TimeControl cannot schedule MATLAB Analysis more often than every five minutes, and channel visualizations update after ten minutes. Platforms listed are API and web.