Date: Sun, 23 Oct 2011 08:09:19 -0700 From: perryh@pluto.rain.com To: kob6558@gmail.com Cc: freebsd-stable@freebsd.org, freebsd@jdc.parodius.com Subject: Re: 8.1 xl + dual-speed Netgear hub = yoyo Message-ID: <4ea42e1f.jwGZJ3vXBxYqD41h%perryh@pluto.rain.com> In-Reply-To: <CAN6yY1vjHHT%2BYoqO6DQPytO-SectkM9V3pJN=uKYHgv26h5LDQ@mail.gmail.com> References: <4ea18916.IBAr0lF5RCzEYn6G%perryh@pluto.rain.com> <4EA12D26.4020104@my.gd> <4ea2b29c.QmX94UmzdHW1HSBe%perryh@pluto.rain.com> <20111022052952.GA2371@icarus.home.lan> <4ea2e874.XnQpdCknhYCB39Py%perryh@pluto.rain.com> <CAN6yY1vjHHT%2BYoqO6DQPytO-SectkM9V3pJN=uKYHgv26h5LDQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Kevin Oberman <kob6558@gmail.com> wrote: > Wow. it's 1985 again. O remember those 10/100 hubs. They were a > royal pain! > > If I remember right, they kept costs down by building in half of > a switch. Traffic from a 10 port to a 100 port was buffered, Speed conversion had to have been buffered in both directions. If trying to convert 100 to 10 unbuffered, only 10% of a packet could have been retransmitted (at 10) by the time the entire packet had been received (at 100). In the other direction, trying to send at 100 while receiving at 10 would have had 9 bit-times of gap for every live bit sent. > but there was no forwarding table and all packets were forwarded > to all ports. I always figured that's normal for a "hub" as opposed to a "switch". > I also remember that SOME hubs of that era had series problems if > the cable was too short. > > You mentioned using a short cable. Have you tried a longer one? > I seem to recall that 3 meters was the minimum, but it was so > long ago that my memory is a bit fuzzy. The first cable I used was about 1 meter, the second less than 2. I suppose I could try a longer one. The only minimum length restriction I _remember_ with Ethernet was the separation of transceivers on a 10Base-5 backbone cable (and, perhaps, a corresponding separation between tees on 10Base-2). I think it was to avoid having two impedance "bumps" within one bit propagation time of each other or some such.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4ea42e1f.jwGZJ3vXBxYqD41h%perryh>