I am wondering if there is any reason for the 568A and 568B standards
for Cat5 network cabling. For instance, would taking any given random
set of wires, matched at both ends, function just as well as a
straight-through cable?

The reason I ask was because I switch my two 24-port hubs over to
24-port switches on a network that I have set up. During that
transition there were two computers that could no longer access the
internet. Up until this point, the cabling from the walls to the
computers were using random non-standard straight-through cabling, as
well as the cabling from the patch panels to the hub. The cabling from
patch panel to the wall was all standard straight-through cabling. I
tried three different patch cables that I made myself using
non-standard straight-through cables, and none seemed to work. I
finally made a cable that follows the 586A standard and suddenly both
computers were able to access the internet from the switch, while all
these cables seem to work fine when connected to the hub.

I suppose this is a two part question, but I hope somebody can give me
some answers. Thanks.

568A and 568B difference is just this :  Transmission and reception
lines are interchanged.

When you connect two computers, A and B, the transmission lines in A
must coincide with reception lines in B and vice versa.  So, you use
568A in computer A side, and 568B in computer B side.

You use straight-through cable when you want an extension only. (568B
in both sides works alike).  You can even create your own standard for
straight-through cables, if you use the same order in both ends it
will work ok.  Of course, that level of creativity is not seen very
professional :)

Here is a nice diagram of the connectors.

http://home.earthlink.net/~bn_dn/cat5.htm

By the way.  The ?A? and ?B? letters are legacy from telephone
industry, where two communicating devices are always called point A
and the point B, which have both transmission and reception
capabilities, just like a phones.

"You can even create your own standard for
straight-through cables, if you use the same order in both ends it
will work ok."

That's what I thought...and changing the cabling is the only thing
that fixed my problem above. So if it wasn't the cabling, what other
explaination might you offer?

Ah !  That is an interesting physical issue.

The standard and female connectors expects that pins 1,2 and 3,6 are
pairs.  As you see in the cable, wires 1 and 2 (green/white ? green)
loops one each other, as well as 3, 6 (orange/white ? orange).

Is there any difference?  Anyway, 1 goes to 1, 2 goes to 2, 3 goes to
3? etc.  It is not important the path that wire takes, right?  Well?
in DC (Direct Current), it doesn?t matter at all, but in hi frequency
it becomes an issue.

You are transmitting bits at rates of 100 Megabits per second.  Those
signals pulling up and down so fast in a thin wire generate a variable
electric filed, which in the meantime creates magnetic fields around
the plastic cover.  The problem is that variable magnetic field
produce an electrical current in the neighbor wires in the opposite
direction!  That means loose of data.  If  ?noise current? is very
high, connection can?t be established.  Even if the connection can be
established, it cause lost packages and slow down the connection.

Well, how the engineers solve that long ago?  They put opposites poles
together.  So, the white/green-white generates circular magnetic
fields clockwise along the cable, but the green wire just at his side
generates the same intensity magnetic field counterclockwise (its
current goes always opposite to white/green), so the magnetic fields
kills each other and no ?noise current? are generated (at least, very
little).

So, if you connect pin 3 with blue wire and pin 6 with white/orange,
their opposites currents may be so far one from the other that their
magnetic fields can?t be cancelled and it may generate noise trough
the wires 1 or 2.

That is the reason for the non-standard cabling mysterious behavior. 
Even when you think it?s working, it may be slowing down your network
speed.

In my times, a ?creative? non standard straight-through worked fine
between a hub and a near computer into a 10BaseT network.  With
current speeds, I don?t recommend that anymore.

Hope I could help.

While I *think* I understand the theory behind your explaination, I
still need some clarification.

If I understand you right, in the case of T568A standard, when looking
at the website you posted ( http://home.earthlink.net/~bn_dn/cat5.htm
), the four wires that actually are used include wires 1, 2, 3 and 6.
The pairs are 1 and 2, and 3 and 6. Now since you want to use the (-)
and (+) of each pair in a twist (i.e. using blue/blue white and
orange/orange white), this is done in both standards. But why are the
RX+ and RX- signals separated in their pin layout by another pair,
unlike the TX+ and TX- pair that are kept next to each other?

The EIA/TIA-568 standard was released in 1991.  It was chosen that way
because the pin/pair assignments in that configuration were compatible
with a wide variety of existing 2-pair voice and data applications.
This configuration has pairs one and two located on the center four
pins, making it compatible with most voice systems, as well as data
systems that used the old USOC pair configuration for 6 conductor and
8 conductor jacks.

Think that you have a single pair phone jack.  Later, you got a second
phone line, so you need a jack that supports two pairs.  What is more
practical?  Doing a new male/female connectors and through away your
old single line jacks?  No, there is more practical to make a female
connector with the first pair in the center (compatible with single
pair jack) and the second pair aside.  So, my female connector works
for both, single pair and two pair male connectors.  That is the
historical reason for the pin assignment in standard.

edullega, thank you so much. You have answered all my questions. Many thanks!

I'm glad to help you.

Ey, there is still the message "There is no answer at this time" for
this question.  ( ? )