A unified query platform for NOSQL databases using polyglot persistence

This research endeavor applies Design Science Research as its principle research strategy as it focuses on the development of an experimental artifact for a unified query system. The artifact encompasses a set of architectural guidelines and principles when a applying a unified querying mechanism for the four types of NoSQL categories: key-value, document, graph and column store data models. The scope of this study is limit to specific vendor implementations, namely: Redis, MongoDB, Neo4j and Cassandra.Ethical Clearance no: 202028917/2023/20A variety of experiments were conducted to evaluate the prototype’s effectiveness and efficiency. The experiments were actioned by a group of automated participants, each test representing a subset of a particular goal. The culmination of these results indicated the feasibility of the proposed solution. The datasets for this study comprises of metrics such as Apdex, error rate, CPU and memory utilization as well as the respective NoSQL generated queries for each data store. The observed data is indicative of how efficient the prototype consumed resources whilst effectively generating an executable query at runtime.<br><br><br><br><br><br><br>本研究以设计科学研究（Design Science Research）为核心研究策略，聚焦于统一查询系统的实验性原型制品开发。该制品针对键值（key-value）、文档（document）、图（graph）与列存储（column store）四类NoSQL数据模型，涵盖了应用统一查询机制所需的一系列架构规范与设计原则。 本研究的研究范围限定于特定厂商的实现方案，即Redis、MongoDB、Neo4j与Cassandra。伦理审查编号：202028917/2023/20A。 本研究通过多组实验评估该原型系统的有效性与运行效率：实验由自动化测试主体执行，每项测试对应特定目标的一个子集。实验结果综合验证了所提方案的可行性。 本研究的数据集包含以下指标：Apdex、错误率、CPU与内存使用率，以及各数据存储生成的对应NoSQL查询语句。观测数据可用于反映该原型系统在运行时生成可执行查询的同时，对系统资源的消耗效率。