Towards a Multidimensional NOSQL DBMS Based on Dynamic and Scalable Data Distribution

Abstract

This work intro duces a new paradigm for Multi-Dimensional NoSQL, namely MDNOS, which targets efficient data management in IoT, Fog, and cloud platforms. It prop oses an overall scalable and fault-tolerant architecture that includes routing in IoT clients. The fog layer namely KV-MDSS, and a cloud layer, namely SD-PGSQL. Dynamic partitioning techniques are integrated into the system for load balance and query p erformance optimization; for single-key data, RP*-SD2DS is adopted, and ZK-RP* for multi-dimensional data. It gives an overview of the basics of SDDS and then go es further to show some realizations of thes e efficiently using MD-NOS. I t presents two sp ecialized architectures: RP*-SD2DS, optimized for one-dimensional data, and ZK-RP*, which extends it further for multi-dimensional data. Exp erimental results prove the supremacy in p erformance and scalability of the system when compared to traditional NoSQL databases like MongoDB with large files and complex queries. It gives reason to b elieve that MD-NOS can enable a revolution in the management of data.