from 13:15 pm to 14:00 pm
Stratio presented its open source Lucene-based implementation of Cassandra’s secondary indexes at Cassandra Summit London 2014, which provided several search engine features. It used to be distributed as a fork of Apache Cassandra, which was a huge problem both for users and maintainers. Nowadays, due to some changes introduced at C* 2.1.6, we are proud to announce that it has become a plugin that can be attached to the official Apache Cassandra.
With the plugin we have been able to provide C* with geospatial capabilities, making it possible to index geographical positions and perform bounding box and radial distance queries. This is achieved through Lucene’s geospatial module.
Another feature we have provided with our plugin is the possibility of indexing bitemporal data models, which distinguish between system time and business time. This way, it is possible to make queries over C* such as “give me what system thought in a certain instant about what happened in another instant”. The implementation has been performed combining range prefix trees with the 4R-Tree approach exposed by Bliujūtė et al.
Both full-text, geospatial and bitemporal queries can be combined with Apache Spark to avoid systematic full-scan, dramatically reducing the amount of data to be processed.
StratioBig Data Architect