Hacker News new | ask | show | jobs
by foota 1251 days ago
I don't think it's the throughput, but the variance. There are so many places for something to go wrong between two computers, so doing something that needs a very tight latency budget is hard. Harder still when you don't have control over any of the parts connecting things. In a DC, everything can be built for qos and latency, but on a desktop you're using a porbably garbage router with a home OS running a million things.