[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: DATACENTER: Pictures of practical, neat, and scalable cable systems



 > We have this with cat three cabling.  Basically we have 25 pair trunks
 > going to various general places in the machine room from 66 or 110 blocks

    We thought of this, but we have to do periodic equipment
rearrangement.  Thus connectorizing the 25pair cable with a CAT5
RJ21X.

 > I'd be interested in seeing the results from this.  I've always been a
 > little wary of mixing signal types on wires in close proximity.  I don't
 > have any good justification for this though and I could be totally wrong.
 > I've also wondered about problems like crosstalk on things like 25 pair

    I was very wary of this as well.  Suffering lossage because of
your cabling is not something I can afford the time for.  In real
world vaguely pathelogical conditions, I couldn't make anything bad
happen.  I really couldn't get the 100fdx ether to hiccup until I put
a 200mhz pulse generator on a neighboring pair in the bundle. :-)

    In theory, I would be very leery of mixing RS232 and other
signaling on the same bundle, since RS232 is single ended, but in
practice, nothing cared...

 > around with is if this would work with 100BaseTX and Cat 5.  Something
 > tells me that there would be some sort of cross talk issues, even if it

    We have a whole floor of a building wired like this, when they
were skimping on cat5 costs and didn't care about T568 compliance.  We
have yet to document lossage from this.

 > was one 100BaseTX with one 10BaseT.  But I'm not an Electrical Engineer or
 > anything so I can't really justify this thought either.

    Nor am I.  YMMV; I've had this box's console and RS232 over pieces
of our cabling plant for the last year, and I haven't noticed lossage
yet.

    BTW, for folks who've been asking me for the URL of my documents;
I'll get it up in the next day or so.  I've got to rewrite what I have
to be a bit more general.  Check my home page, a link will appear in
the next day or so.  (http://www.ai.mit.edu/people/hag/)