Simulating Multi-Model Data Evolution for Benchmarking Big Data Systems

This paper addresses the challenge of benchmarking multi-model data management systems capable of handling diverse and evolving data. Existing benchmarks are typically static, limited to specific models, and insufficient for evaluating cross-model interoperability or schema evolution. To overcome these limitations, we introduce a novel tool, called TransforMMer, that enables the generation of dynamic, customizable benchmarks from heterogeneous datasets. The tool combines schema inference, editing, transformation, and export within a unified graphical interface. It supports multiple data models and schema versions, facilitating comparative performance evaluation across systems. Experimental results on real-world datasets demonstrate its effectiveness and adaptability. To promote reproducibility and community adoption, we also provide DaRe, a curated repository of benchmark datasets.