[j-nsp] non-concatenated STM-4 on Junipers

[Dmitri Kalintsev]

On Mon, Feb 10, 2003 at 08:25:11PM +0100, Daniel Roesen wrote:

[cut]

> BTW... this also happened because the carrier wasn't able to
> deliver STM4c with his equipment. :-)

Ugh, STM4c vs STM4. Now that you've mentioned it, I do recall seeing a "c"
notion before but have never paid attention to it (serves me right, and
*now* I know what "VC4" is, too. ;^) ). This has bitten us in the bum also,
and we were extremely surprised to find out that our SDH vendor (one of the
biggest ones there are) had to *manufacture* concatenated STM4 cards
especially for our deployment.

> > If given four STM-1 channels, I suppose we could configure a "SONET
> > aggregated interface" in the software and re-bundle the channels
> > logically,
> Yep, but only starting in JunOS 5.6.

Have you tried this, by the way? I seem to recall that logical concatenation
is done using Juniper's load sharing, and if you don't have enough FECs for
it to work well enough, you may never get to the true capacity of the full
STM4.

SY,
-- 
D.K.