The Semantic Data Stream Manager provides a middleware layer for the intregration of data streams with semantic web technologies. In particular it provides an abstraction for the semantic represention of streams or arbitrary content, and tight integration with Tupelo to store and query streams using triple stores.
Features include:
- Timestamping of RDF resources in a time series to represent streams, both for RDF graphs and data blobs.
- Blobs are stored in a user-defined, dynamically-set granularity to provide more control over read/write times.
- Transparently manages missing data points in a time series, points seen out of sequence, varying frequencies, etc.
- Integration with Tupelo allows the use of several RDF stores and other s storage frameworks.
- Java API.
- Capturing and publishing data from/to Open Source DataTurbine.
- A SPARQL extension to query streams.
- Tools for use in Cyberintegrator workflows, fetching streams via http, ftp and pop, publishing streams via http.