HDFS usage for small files

Question

Suppose we have two situations and they are having a small amount of data distributed across different files and a large amount of data in one file. In which situation we should use HDFS?
Little confused over here

Answer 1

Well, the HDFS works more reliably with the large data when the same is stored on a single file. In Name Node, the concerned information is available in the RAM and thus it cannot deal with a large number of files. In case files are more, there will be a large amount of Metadata it needs to deal with. It is almost impossible to store such a large volume of Metadata in the RAM.
Hope this clears everything