NLA-Bit: A Basic Structure for Storing Big Data with Complexity O(1)

Abstract

This paper introduces a novel approach for storing Resource Description Framework (RDF) data based on the possibilities of Natural Language Addressing (NLA) and on a special NLA basic structure for storing Big Data, called “NLA-bit”, which is aimed to support middle-size or large distributed RDF triple or quadruple stores with time complexity O(1). The main idea of NLA is to use letter codes as coordinates (addresses) for data storing. This avoids indexing and provides high-speed direct access to the data with time complexity O(1). NLA-bit is a structured set of all RDF instances with the same “Subject”. An example based on a document system, where every document is stored as NLA-bit, which contains all data connected to it by metadata links, is discussed. The NLA-bits open up a wide field for research and practical implementations in the field of large databases with dynamic semi-structured data (Big Data). Important advantages of the approach are as follow: (1) The reduction of the amount of occupied memory due to the complete absence of additional indexes, absolute addresses, pointers, and additional files; (2) reduction of processing time due to the complete lack of demand—the data are stored/extracted to/from a direct address.