dnwcv@dcatla.UUCP (William C. VerSteeg) (03/22/88)
I have a configuration that requires a two ethernet based devices to communicate with each other. Not too uncommon (:-)). But, I know that these two devices are the only two devices that will ever be on this ethernet segment. So, I wonder, why do I need the ethernet segment and two transceivers ? I could use a low cost LNI, but even that has capabilities that I don't REALLY need. Is there a way to cable the output of the first device to the input of the second device such that no active components are needed? Failing no active components, can it be done in a way that the devices can be connected with a low cost device that is powered by the power normally used by the transceivers? If such a device is possible, is it commercially available? If it is not available, has anyone seen work done in this area ? Thanks in Advance Bill VerSteeg
kwe@bu-cs.BU.EDU (kwe@bu-it.bu.edu (Kent W. England)) (03/23/88)
In article <3071@dcatla.UUCP> dnwcv@dcatla.UUCP (William C. VerSteeg) writes: > >So, I wonder, why do I need the ethernet segment and >two transceivers ? I could use a low cost LNI, but even that has capabilities >that I don't REALLY need. Is there a way to cable the output of the first >device to the input of the second device such that no active components >are needed? This is the famous request for a null transceiver cable. In answer to your question, as with a null RS-232 cable there are certain signals that can't be duplicated with a null cable. You could swap xmit and receive, but you couldn't reproduce collision detect or heartbeat. Heartbeat is no big deal, but what about collision detect? For only two stations, why not go CSMA w/o CD?? This question has been asked often enough, has anyone actually tried doing this? I haven't, but I'm getting curious. I might get curious enough to do it. But before I do has anyone tried a null transceiver cable with only xmit and rcv swapped? How'd it work? Kent (in a silly mood) England Boston University [If you can run drop cables off Ts with thin-net, why not null xcvr cables?]
jqj@uoregon.UUCP (JQ Johnson) (04/01/88)
>This is the famous request for a null transceiver cable
American Network Connections, (408)-737-1511, advertises a
transceiver-like device, their ANC-10 Direct Connection Device, that
would seem to be exactly what is requested (including CD and SQE). I
have no experience with the company (it seems to be a small startup),
and don't even know the price of the device (probably about $200, since
a pair of transceivers costs ~ $300). ANC also advertises a 2-port
Ethernet transceiver and a DELNI-equivalent.
If anyone has actually tried their products, I'd be interested in a
review.