old modems & a packet switcher

From: Frank McConnell <fmc_at_reanimators.org>
Date: Mon Aug 3 13:55:08 1998

William Donzelli <william_at_ans.net> wrote:
> Lots of the stuff I have seen allows a user to insert an external clock -
> this can be handy for testing in the real world. Providing the clock is no
> big deal. Sure, it would be nice to get a dead-on clock for a T1, but as
> long as you keep the circuit private, the clock can be fairly sloppy.

Thanks for jogging my memory about that, I guess I've been in denial
about this stuff for a while. Yeah, if you're just running over a direct
cable (which is what I was thinking about) you'd be OK doing this.

> > Can you do this over dry copper, for short runs?
>
> Yes, some people on this list get connected this way (T1s that run a whole
> 20 feet, CSU to CSU, with no frame in between). A horrible kludge in a
> production environment, in my opinion, but it worked for quite a few
> years. Recently someone at ANS suggested that we use a T3 the same way,
> but he ended up being tarred and feathered.

Umm...why were you doing this? I can understand it for testing and as
a crude but expensive bandwidth limiter.

> > If so, would you want to? You mentioned in another post that you'd
> > have to run conduit for FDDI, and that makes me think there might be
> > some open space between hither and yon. Sure, you could run copper,
> > but wouldn't electrical potential differences be a problem?
>
> Differences as in losses due to the long run? That is what the Line Build
> Out option is for (on most CSUs). T1s, with thier twisted pair,
> differential design, tend to be pretty tough.

No, I was thinking about differences in ground potential between the
two buildings, static electricity, lightning strikes, that sort of
thing. All the canonical reasons why you wouldn't want to pull
Ethernet coax through that conduit.

-Frank McConnell
Received on Mon Aug 03 1998 - 13:55:08 BST

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:30:43 BST