From: Jay Hennigan (jay@xxxxxxxx)
Date: Fri May 18 2001 - 18:00:16 GMT-3
On Fri, 18 May 2001, Jeongwoo Park wrote:
> Hey jason
> Thanks for your input.
> Why should it be set to network or line rather than internal? is it line or
> DTE by the way?
> the csu that I have has three option for clock source: line, dte, and
> internal.
> Can you tell me the brief difference of those?
Line is the same as network. The CSU will recover clock from the T-1 line
or network, use that recovered clock to transmit back to the line, and also
use that recovered clock to synchronize to the router. At least one of
your CSUs wants to be set this way. Here's where it can get tricky in the
real world.
Internal means that this CSU itself will generate clock, and use this
internally generated clock to drive both the T-1 line (network) and the
router. One (and only one) of your CSUs _may_ want to be set this way.
DTE means that the CSU takes its clock from the router which must then
have a DCE cable and the "clock rate" specified. It is very unlikely
that you'll have a real-world T-1 set up this way.
So, in the real world, you'll either have one CSU on Internal and one
on Line, or both on Line. Let's look at two possible setups:
Setup 1: Two CSUs connected by a T-1 line that's just a couple of
pairs of wires. This can be used in a lab setup in place of the DTE/DCE
cables, or for testing with both routers next to each other. In this
setup, if both CSUs are expecting the "line" to provide a clocking
signal, then there's an obvious problem, as the "line" is just copper
wire. Hence, no clock signal and things don't work. Therefor you want
one and only one CSU to use an internal clock. The other CSU is set to
"Line", and happily uses the clock it gets over the wire to synchronize
itself and its connected router.
In the real world, *some* telco circuits are indeed provisioned this
way, and the point-to-point T-1 is just wire. These are typically shorter
circuits within a single city.
Setup 2: The T-1 line between the two ends is provisioned on a
facility that has clocking provided by the carrier. This is usually
true if the circuit travels a long distance, as it's easier for a
telco to multiplex T-1s onto a larger pipe that has clocking from
a stable central source. In this case, if either of your CSUs is
set to "Internal", it will ignore the clocking that telco is providing
and try to synchronize everything to itself. Because telco is also
providing a clock, it will ignore the clock provided by the CSU.
This results in what are called "slips", which are very destructive
to data. Over time the two clocks get out of step, then come back
in tune, then drift apart. This causes errors, similar to waht you're
seeing.
Hopefully, the carrier from which you ordered the circuit can tell you
whether or not they are providing clocking, and the linecode (B8ZS or AMI)
and framing (ESF or SF/D4). You'll need to set your CSUs accordingly.
-- Jay Hennigan - Network Administration - jay@west.net NetLojix Communications, Inc. NASDAQ: NETX - http://www.netlojix.com/ WestNet: Connecting you to the planet. 805 884-6323 **Please read:http://www.groupstudy.com/list/posting.html
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:30:45 GMT-3