[j-nsp] LX SFP Question

Richard A Steenbergen ras at e-gerbil.net
Fri Sep 25 09:59:24 EDT 2009

On Thu, Sep 24, 2009 at 05:35:20PM -0400, Paul Stewart wrote:
> Hi folks...
> Does anyone know the tolerance of the LH SFP's from Juniper?  We are trying
> to get an EX3200 switch configured and ready for production - have a case
> open at JTAC but haven't been able to resolve.  In fairness to the JTAC
> engineer, I haven't had a lot of time to troubleshoot except for performing
> a software upgrade which has been completed (9.4)
> The link is up/up from the EX3200 to a Cisco 6500 but the distance at the
> moment (while testing) is literally 15' or so.  In the Cisco world we have
> no problem on such short distances but wondering if something is different
> or causing a problem for the Juniper.

Juniper LH != Cisco LH. You're probably blinding your optics, you neeed 
to change them or add an attenuator for such short distances.

Juniper LH == 1550nm with 70km reach, equivalent to Cisco ZX (well not
exactly equivalent, Cisco ZX optics have a 100km power budget, but the
optic type is exactly the same just with less receive sensitivity).

Cisco LH == 1310nm with 10km reach, equivalent to Juniper LX.

I believe the confusion stemmed from the fact that LX isn't actually
10km like everyone thinks it is. The original 1GE spec called for LX to
have a reach of 2km, but when it actually got implemented essentially
everyone ended up producing a 10km version so nobody even remembers the
2km version. Cisco called this 10km version "LH", and writes "LX/LH" on
all of their 10km optics to indicate that they are the same thing just
with a better optical budget. Don't ask me why Juniper decided to go
with the LH name for their 1550nm optics, but it's been confusing people
and blinding receivers for the last 10 years or so. :)

Richard A Steenbergen <ras at e-gerbil.net>       http://www.e-gerbil.net/ras
GPG Key ID: 0xF8B12CBC (7535 7F59 8204 ED1F CC1C 53AF 4C41 5ECA F8B1 2CBC)

More information about the juniper-nsp mailing list