Good point. GFP, by definition in ITU-T (G.7041/Y.1303 ), was designed to operate over any octet synchronous transport network, which would include G.709 digital wrapper. Looking through the standard, however, I don't see anything special in regard to G.709. SDH and OTN are used synonymously. Someone from Lucent would be best to comment here, as they remain the only large vendor really pushing OTN forward.
re: GFP: Good For PositioningIn answer to your post, the real beauty of RPR is that it can be deployed as a blade into an existing vendors' (Nortel, Alcatel, etc...) box. Obviously the SONET/SDH ring has to have unused capacity (if it didn't then you'd be in the "overbuild" scenario you say is not the case), and you would then add an RPR blade for any data traffic you wanted to carry IN PARALLEL with your existing TDM traffic.
True.
But RPR delivers little benefits over a traditional SONET/SDH box that has a native Ethernet switching blade with GFP/VC/LCAS as the 'encapsulation' and UPSR/SNCP on the line side.
the two were merged along with contributions from others (ie: Luminous) into the RPR protocols, that are available today as a standard blade off nearly all SONET/SDH vendor's boxes
Hardly. I can only think of Nortel. Even Cisco does not do it with their '545/'327. Why don't you name others?
On the otherhand, there are numerous vendors who implement simple native Ethernet switching blades with GFP encap.
re: GFP: Good For PositioningScott Clavenna wrote...
"Again, the drawback here seems to be the need to deploy digital wrapper all through the network. When using GFP-over-SONET/SDH, the core network remains untouched. A big plus in this market."
Well at one point I was fairly well-known in some circles for saying "Digital wrapper? We already have a digital wrapper. It's called SONET", so I'm not a HUGE fan of G.709 (although there are some very elegant things about it), so I think your comments are pretty right on. (The market's already spoken on the issue, methinks.)
BUT, at 40Gig, we're going to need some very serious FEC capabilities, which G.709 has in droves. So G.709 plus GFP stands a small chance of getting used at those rates.
SO there's actually a carrier that has deployed G.709 end to end? Well, wonders never cease! (The again, an ITU standard has a lot more meaning in Europe I think.)
re: GFP: Good For Positioningstraight shooter writes: "... In the end though, whether RPR or GFP, both are just "blade" enhancements and are constrained by SONET/SDH envelopes. Try to create a 20K endpoint VPN/TLS offering as are currently demanded and/or in deployment in Asia and the SONET/SDH solution breaks."
1. GFP is not an alternative to, but in fact part of the IEEE 802.17 draft for RPR. (GFP can of course also be used in other ways, independent of 802.17 or similar proprietary approaches.
2. If you find SONET/SDH envelopes too "constraining" at only up to 10Gbps capacity each, then build whatever higher capacity interfaces your customers require and sell as many as you can. (Meanwhile, in the real world, work on VT1.5 virtual concatenation will continue.)
3. The number of endpoints has nothing to do with the capacity of the transport network.
re: GFP: Good For PositioningOSXman writes: "What are the advantages of low order vc and where would that be implemented (vs. high order vc)?"
High order (e.g. STS1/VC3/STS3c/VC4) is good when you need more than 50Mbps of capacity; Low order (e.g. VT1.5/VC12) is good when you need more than 1.5 or 2 Mbps but less than 50Mbps.
Good point. GFP, by definition in ITU-T (G.7041/Y.1303 ), was designed to operate over any octet synchronous transport network, which would include G.709 digital wrapper. Looking through the standard, however, I don't see anything special in regard to G.709. SDH and OTN are used synonymously. Someone from Lucent would be best to comment here, as they remain the only large vendor really pushing OTN forward.
Scott