TritanDb is a time-series database for Internet of Things Analytics with a rich graph data model and fast lightweight core. It can be used to ingest, store and query time-series data in real-time. TRITAnDb stands for Time-series Rapid Internet of Things Analytics Database.
- Best-of-class compression specifically optimised for time-series data.
- Fast queries and aggregation utilising a specialised immutable data structure called TrTables.
- Fast data ingestion utilising a ring buffer.
- Event sourcing to support journaling.
- Accepts out-of-order events.
- Rich graph data model like RDF are supported.
- Powerful graph query engine with support for SPARQL.
- Optimised for Fog Computing across lightweight Things as well as cloud servers.
- Written in Kotlin and runs on the JVM.
You can build TritanDb using Gradle
git clone https://github.com/eugenesiow/tritandb-kt.git
cd tritandb-kt
./gradlew build
Start the TritanDB server:
./gradlew execute
Get the TritanDB Command Line Client (tritandb-cli) and issue commands:
usage: tritandb-cli [-h] [--ingest]
optional arguments:
-h, --help show this help message and exit
--ingest, mode of operation
--query,
--list