Dremio Join of large table in S3 and RedShift

1)While joining large tables will if there is a data skewness, how dremio address it?
2)Will the join be happening in dremio nodes?
3)If we have zero data copy how these joins work inderhood?

@ann In case of Parquet on S3, joins will happen on Dremio but you can reflect the VDS that has the joins so they are not computed at run time, there will be skew if filter applied causes it, in the next few version the planner will be more efficient in handling data skew