It seems to me that Lombrozo is making a mountain out of molehill when he talks about latency. We're not talking about transferring a block that is 10GB here.
Btw, are there any research papers addressing the latency vs. block size "issue?"
You are viewing a single comment's thread from: