A NoSQL-Based Approach for Data Resource Allocation Problems: Embedded Systems Use Case^*

Abstract

To allow further automatic big data management, and to reduce data resource allocation problems, we propose an approach for embedded systems reconfiguration. The reconfiguration of embedded systems consists of adding new servers and clusters for exchanging the big data. For that, we propose a methodology based on an automatic programming approach to deploy new data servers based on NoSQL performances. These servers will be interlinked to get generated data from an embedded system. This will increase the number of satisfied transactions and queries with an optimal data access latency. Moreover, to evaluate the performance of our model, we propose a new definition based on committing full queries and the ability of all the servers to manage and process the generated big data with the constraint of time. To test and validate the proposed approach, a software tool named SFRBDM (Software for Real-time Big Data Management) has been developed. This software tool allows users to simulate the functioning of embedded systems, inject, generate data, and ensure recoveries between data servers and clusters. SFRBDM also allows calculating a better resource allocation of the simulated embedded databases with a higher performance based on NoSQL. The experimental study performed over multiple servers, with different structures and sizes, validates the efficiency of our proposed approach.