Heterogeneous NoSQL Databases Abstraction Approach Based on Full Text Search Indexes

Abstract

The exponential growth of unstructured data in the mobile applications, the social networks and the web technologies led to NoSQL database emergence. While this specific class of DBMS provided a better scalability for databases, the lack of a standard DML that unifies and simplifies querying NoSQL data stores is still a hard deal especially in heterogeneous environments. A simple SQL query can turn into a complex map-reduce function in the NoSQL world in order to obtain the same result in the standard SQL DDBMS. With no common convention between the large variety of NoSQL implementations and families, each product implemented its vision of the NoSQL concept. Each implementation covered distinct functional scopes, depending on the target domain and the creation purposes. Meanwhile, many successful NoSQL databases integrated a powerful full text component to enhance their search capabilities. To remedy this variety limitation, we propose a new incremental approach that allows (1) the standardization of NOSQL search queries among heterogeneous NoSQL data stores and (2) NoSQL search queries optimizing. This approach is based on (1) the definition of a new universal engine for full text indexing, (2) incremental synchronization of data and indexes between the stretched sites.