[180656] in North American Network Operators' Group
RE: 100G DWDM FEC standard
daemon@ATHENA.MIT.EDU (Peter Lothberg)
Tue Jun 9 10:47:02 2015
X-Original-To: nanog@nanog.org
Date: Tue, 9 Jun 2015 16:46:45 CEST
From: Peter Lothberg <roll@Stupi.SE>
To: Gert Grammel <ggrammel@juniper.net>
In-Reply-To: <BN1PR05MB0416461AFEF1AEDEA25449ECEBE0@BN1PR05MB041.namprd05.prod.outlook.com>
Cc: "nanog@nanog.org" <nanog@nanog.org>
Errors-To: nanog-bounces@nanog.org
> From a standards perspective keep in mind that
> http://www.stupi.se/Standards/100G-long-haul4.pdf is not approved -
> but we are working hard on it. OTOH having a reference
> implementation at hand, is an accelerator that helps a lot.
There is a whole industry that do not want it to be plug and play...
(The ones that do not make routers or switches..)
> Let me also add some color to your email as the current
> interoperability situation in WDM is quite funny. Sometimes
> transceivers of the same vendor can't talk to each other, as they
> are based on a different generation of ASICs and therefore FEC
> implementations. In other words, vendors typically have more than
> only one secret sauce they cook with, and different sauces do not
> blend well :-) . Perhaps transport folks are already too used to
> deal with such kind of issues that no one laments anymore. On the
> other hand perhaps, the networking industry is already so used to
> Ethernet where interopera bility is a no-brainer, that it is
> difficult to imagine what it means to deal with a technology that
> prevents multi-vendor interop.
All vendors have secret souce for 100G SD-FEC, and just the fact that
you can wire the wire the differential encoding eight ways..
That;s why we settled on a HD-FEC that can be inside the DSP-Asic or
inline after it. All to us known DSP implementations supports this
with more, less or no extra logic. And the logic is free and fully
documented.
> To confirm your final point:
> Interoperability is on the top of the shopping lists for the
> networking industry.
Amen!
-P