[c-nsp] Re: T1 circuits over OC3
andy at xecu.net
Fri Oct 1 00:30:24 EDT 2004
On Thu, 30 Sep 2004, Jay Ford wrote:
> On Thu, 30 Sep 2004, Andy Dills wrote:
> > If only you could hand the telco equipment and say, "Use this." :)
> This gear goes on the outside edge facing the user of the T1, basically
> delivering a T1 through an ATM cloud. All the telco has to provide is a PVC,
> ideally (& perhaps required, I don't remember) CBR. My understanding of the
> original request was for T1 through an ATM cloud, so this seems to fit. On
> the other hand, I've never touched telco frame service so I could be off base
I don't think they wanted to use an ATM T1, as nobody wants the cell loss
on low speed links like that. And few people have that type of CPE
equipment laying around.
See, I thought the requirement was for terminating a traditional leased
line, for instance using HDLC or PPP, but just accepting the T1 over the
ATM. Similar to a channelized DS3.
And my answer was that something needs to translate on L2. That's where
FRASI comes in. The customer configures their T1 for frame relay, and it's
delivered to the service provider as rfc 1483.
However, and this is important (and something I pointed out in the
original email), in order to utilize FRASI, you _must_ police the cells
you are sending them. I have only seen this done using vbr-nrt. If you
don't police the cell rate, it will not work anywhere approaching
> > I'm curious, what is the encapsulation of the T1 circuit between the LS
> > and the CPE?
> Each T1 circuit emulation card has 4 RJ-45 jacks pinned as RJ-48c for a plain
> old 4-wire T1. They do the usual T1 framing (ESF...) & encoding (B8ZS...).
> Is that what you were asking?
That's L1, I'm talking L2.
More information about the cisco-nsp