"Run Batch Simulations Without Recompiling Generated Code" with Big Data

How do you run batch simulations with big data? Can you pass in a dataset reference MAT file?

The example here, https://www.mathworks.com/help/rtw/ug/run-batch-simulations-without-recompiling-generated-code-e251b63f1b75.html, is where I started. It's great, and I have benefited from it. However, a simulation will fail if the input dataset is too large. Can I merge this example with big data techniques? Does an example exist somewhere?