Space Curve Technology


Tackling a new generation of analytical challenges.

Space Curve’s platform is unlike any other distributed analytical database. It is not based on Hadoop or distributed hash tables. The algorithms and data structures were developed for large-scale database applications like real-time geospatial and graph analysis that traditional big data technologies handle poorly.

Our technology originates in research aimed at solving the interval indexing algorithm problem for Google Earth. Put simply, how do you design a distributable indexing algorithm for a dynamic set of overlapping polygons? This has been a persistent unsolved algorithm problem for parallel databases. A general solution to the interval indexing problem demonstrated in 2007 evolved into a new model for designing distributable algorithms. The underlying computer science has since been extended to address additional algorithm problems including the notoriously non-distributable self-joins that are central to graph analysis.

Several aspects of the architectural model are unusual. Operands for the underlying algebra are interval data types. The overlay network efficiently addresses a rich range of distributed operation, including selection and join operators. Sharding functions are based on a novel value-agnostic similarity measure. Data is organized by a logical embedding in a hyper-dimensional topology. Though more abstracted and expressive, the computational overhead of the representation is similar to other distributed data structures.

Space Curve’s technology is the foundation of a new generation of big data platforms allowing faster and deeper insights than ever before.

Build a living model of the world.

The system architecture is designed from inception to construct a precise mirror of reality at the speed of life. Geospatial data models are seamlessly distributed, real-time, and fully geodetic by default. Geometry operations and spatial relationships accurately reflect the curvature of the earth. Unlike many projection-based geospatial systems, there are no restrictions on polygons overlapping poles or meridians. It represents the world the way we experience it.

Track millions of objects in real-time, fusing machine-generated and human-generated data into single data models. Complex data such as geospatial polygons are spatially indexed at sustained rates of millions per second on commodity server clusters, supporting ingest of sensor-based data sources. Queries are concurrent with data ingest enabling immediate online analysis of new information. Space Curve lets you to analyze reality by measuring it.