One of my spark job is doing a lot of data transfer and I am wondering what could be the way to optimize data transfers within spark in general?
