This paper is well worth reading: http://groups.csail.mit.edu/ana/Publications/PubPDFs/1988Analysis%20TCP%20Processing%20Overhead.pdf
While the traditional BSD implementation uses mbufs that complicate things, actual tcp processing can be done quite cheaply. > On Nov 21, 2014, at 6:34 AM, erik quanstrom <quans...@quanstro.net> wrote: > >> On Thu Nov 20 13:44:04 EST 2014, a...@9srv.net wrote: >> Both. I agree with what you're saying about the computers, but I was >> thinking of the fact that the wire speed is fast enough in most cases that >> the tcp/ip overhead doesn't impact things noticeably for most uses. There >> are outliers in both cases, of course. > > this is not correct. tcp doesn't help at all when the wire is fast (short, > fat). it's the classic tradeoff of cpu > for (networking) performance. the wire being fast enough is an argument > against using tcp, > not for it. > > so really, it's the gobs of cpu we currently have that make tcp not an issue, > not the gobs of bandwidth. > > - erik >
