Positioning CortexDB – Innovative Bi-Temporal, Multimodal NoSQL Technology

Research Letter
March 2015
Positioning CortexDB – Innovative Bi-Temporal, Multimodal NoSQL Technology

By Dr. Wolfgang Martin

CortexDB is a bi-temporal, multimodal NoSQL database technology proving a flexi- ble platform for agile enterprise web applications. It has been developed based on findings in brain research.

What makes CortexDB unique?
It differs from all known databases via its index structure.
Different types of database management system are determined by their database schema. In the case of relational databases, data are organized in tables with columns and rows. Columns are nominated as unique identifiers and represent relationships across tables. Columns with the same domain are used to join tables. Other NoSQL databases (NoSQL = “Not only SQL”) have a different database schema like key-value store, document store, etc. All these databases are different in the way they organize data. Also, in many cases, it is the application that defines the schema. The use of indices is common for all databases. Indices are flat structures of data with a reference to records (either table rows, document IDs etc.) so that they can access data in a sorted order. Hence, special indices for different types of data make up the main differences between databases, but their index structure is always flat.
CortexDB works differently. At first glance, it is a schemaless database, and one may compare it to document stores, but it is much more. It comes with a content based index structure (CorAIT1). This means that each item of content (value) knows in what kind of fields (keys) and in which records (Doc-ID) it exists, and every key knows what different types of values are used. So, CortexDB has a universal index of all fields with all occurrences representing the whole database. Consequently, all data queries are based on the index structure without the need for joins. This also implies that a database schema is only used for output of data (data records or data documents). The memory consumption for all indices is about the same size as the size of a CSV file containing the raw data.

DOWNLOAD Research Letter (.PDF): ResearchLetter_Cortex_180315

You may also like...