[nSLUG] ethernet and the modern linux

Daniel Morrison draker at gmail.com
Fri Sep 26 13:53:28 ADT 2008

2008/9/26 Jason Kenney <jdkenney at gmail.com>:
> Yes... I had wires written originally and decided to switch to pairs
> and forgot to change a number.  Clearly it was a typo.

Sorry man, it was just confusing.

> I don't
> believe it makes my point about gigabit ethernet requiring
> significantly more complexity to operate in ("just") a half-duplex
> mode compared to the 10/100 cases any less valid.

You're absolutely right about that, and I didn't realize before, in
fact, how different they were.

> If you'd like to
> address instead, you know, the point I was making: ie gigabit ethernet
> is so different from 10/100 you shouldn't group them together, then I
> am all ears.

"not group them together" in the sense of making policy decisions for
both 100 and 1000 in the same breath.  Totally in agreement.  "not
group them together" in the sense of mixing 100 and 1000 equipment on
the same network... yeah, better to run parallel networks if you can,
or segregate the 1000 equipment together and the 100 equipment
together.  Make certain the device that links the two together is a
capable one.  No disaster to have a few 1000 pieces of equipment on an
other 100Mb network, but it's a waste.

>> BTW, WIkipedia says:
> Wikipedia should have a program where by reading an article you
> qualify to receive a certificate in the mail that says: "GOLD STAR:
> _________  IS NOW AN EXPERT".

Hey, Wikipedia is a great resource.  Just don't cite it as a reference
in a paper -- go get Wikipedia's own reference and interpret it in
your own words.

Obviously one shouldn't trust anything written in Wikipedia blindly --
but that's a misleading statement.  It can be easily simplified to:
one should never trust anything blindly.  I don't even trust myself
blindly.  That's why I checked Wikipedia to confirm my understand that
1000BT was 4 pairs while 100BT was 2 pairs. I might have been wrong,
you know!


More information about the nSLUG mailing list