On Thu, 26 February 1998 at 0:25:04 +0000, John Kelly wrote:
> On Thu, 26 Feb 1998 08:36:15 +1030, Greg Lehey <firstname.lastname@example.org> wrote:
>>> It doesn't jive at all with test results Tolly published several years
>>> ago in Data Communications.
>> Did he have an axe to grind?
> Not that I know of.
>>> He said token ring would run at full 16mb wire speed while Ethernet
>>> would degrade to 7mb because of collisions.
>> Right, and none of us have ever seen more than 7 Mb/s out of a 10 Mb
>> Ethernet, right?
> You're dodging the collision issue.
Not at all.
> I've seen 10mb between two stations on my BNC Ethernet when nothing
> else is on the wire, but more stations talking will cause throughput
> to fall off.
Sure. That's the whole point. Without stating the test conditions,
the figure of 7 Mb/s is meaningless.
>> Any such statement *must* be qualified by the test conditions.
> He probably tested it more thoroughly than either one of us have.
> Why not read his report instead of throwing stones?
Did you publish a URL? In any case, I wasn't so much complaining
about the report as picking an unqualified value from it.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message