Bruce Lane wrote:
On 13-Aug-06 at 11:26 Don wrote:
So, there is nothing *active* in the cables?
E.g., the
"mainframe" has all of the "adapters" inside (on the
appropriate cards). Does this pose problems if you
plug the wrong cable into a card (i.e. if cable D shell
pins X & Y expect to be tied to a USB mouse and suddenly
find EIA232 signal levels on those pins)?
As far as I know, from a cursory examination, there is nothing
active about the cables. They're just wire.
As for plugging the "wrong" cable into a card, I'm not sure
if that would pose a problem. I think Doc mentioned that each
card is a 'universal' interface which is jumper or DIP-switch
settable for the machine type it's dealing with.
So, if you strapped the "mouse pins" for USB and then
erroneously plugged a cable into the "universal" D connector
that had a SERIAL mouse attached...?
I.e. are there any easy/mindless ways to toast things?
So I can't
(readily) use a Sun keyboard -- unless I add
an external adapter...?
I'm pretty sure not, though I could be wrong. Look at the
bright side: The IBM 'Clicker' keyboards are still pretty nice.
<grin>
<frown>
Amusing to think someone will spend $125+/node
just to live WITHOUT a keyboard! :-(
Oh, you'd be surprised! Look at big companies with big pockets,
like my former employer (Boeing). They would happily spend five
figures on switching equipment to put Lord only knows how many
systems in the data center into one or two KVM combos.
Oh, I've seen many server farms -- but usually they don't
bother with consoles and rely on serial consoles instead
(or other "management" functionality over network).