Re: [tcpm] TCP tuning

[Joe Touch <touch at ISI.EDU>]

rick jones wrote:
...
> My experience thusfar with netperf benchmarking has been that the
> single-transaction at a time TCP_RR test has never run into
> Nagle-induced delays on any stack that wasn't broken, and I cannot
> recall the last time I encountered such a broken stack. 

What you have seen is consistency in implementations. Such behavior is
not required by the specs, so it should not be considered "broken", nor
should tests rely on it.

Thanks for removing it as an issue, though. There are many other things
that could be happening, and as I noted the behavior was never really
well spec'd AFAIR.

Joe

Attachment: signature.asc
Description: OpenPGP digital signature