Compare to Relational Databases

When it comes to databases, one size doesn’t fit all. SciDB is a computational database – not a data warehouse, not a business intelligence database and not a transactional database.

Relational databases were built for handling a lot of transactions. These systems have the wrong data model for basic data exploration where you want to look at what happened within a range (of time, geographic locations, probe locations, etc.). Moreover, these systems do not store data in a format that is convenient for complex analytics—forcing users to move and rearrange data before they can do any linear algebra. Machine-generated data, scientific data and data with spatial or temporal ordering just do not fit neatly into relational tables. While they can be made to do this work, they will be slow and  inefficient in their use of storage.

SciDB’s multi-dimensional array data model is a natural match for complex data like genetic sequencing data, sensor data, financial data and geospatial data. And since SciDB is an all-in-one database and analytics platform, you don’t have to extract, reformat and export your data to a standalone math package.

SciDB is designed from the ground up for large-scale complex analytics:

  • Provides automated, transparent support for massively distributed data sets.
  • Accelerates array operations by 10-100x because data is stored natively in an array format.
  • Natively integrates complex analytics directly with data management, supporting ad-hoc analytics at scale.
  • Supports reproducing and auditing of results because data is never overwritten, only updated.