Understanding the Technological and Economic Issues of FTTH Architecture
February 23rd, 2007 by Gordon Cook
In the debate about fiber to the home one of the significant issues is deciding on the architecture out of these decision fall impacts on cost and on control of the link. Will it be by the carrier, the home owner or the community network? Consider the following exchange on my Economics of IP Networks mail list today.
See Wikipedia on Passive Optical Network.
Bill St Arnaud: PON of any flavor does not make sense. The biggest drawback with PON, whether it is WDM-PON, G-PON, BPON, EPON etc, is the fact that the Optical Network Terminal (ONT) at the customer premises must be under the control and operation of the carrier. This is required to insure that there is compatibility and synchronization in the upstream signal is usually based on a time slice protocol or an assigned wavelength. Almost everybody focuses on the downstream capability and features of PON systems and completely ignores the upstream path. The upstream path is usually orders of magnitude less bandwidth than the downstream path. It also has distance and timing limitations. In new fiber builds the incremental cost to go with home run fiber is trivial compared to distributed PON. Even with existing fiber deployments it is often far cheaper to do an overlash with a new home run fiber cable, rather than deploying PON The only advantage I can see to PON is that it reduces interface count on the central office termination gear. But low cost termination optics from companies like Infinera will soon remove that advantage.
Fiber Network Builder in Europe: I agree with Bill. PON means that over 70 % of the value chain is locked-in (physical layer, transport layer), which drastically reduces the option value I in the economic sense. Doing this makes no sense for infrastructures which are supposed to have an economic life of 30 yrs or more. You hurt your economy’s potential to evolve and grow.
Frank Coluccio: I don’t think anyone here, including myself, is “for” PON. In an earlier post, Bob made a comment suggesting that wavelengths over PON would be preferable, and I merely demonstrated that it was already available.
Bill, most of the negative assertions you’ve made [are] about PON’s asymmetrical characteristics and loss of end user control (”the customer premises must be under the control and operation of the carrier”). [in my opinion these are clearly your own projections of how the technology would be implemented, and not reflections of any inherent constraints imposed by the technology. It can be implemented in the manner you suggest, and it would be bad. Or it could be implemented in myriad other ways, some no different than R-PON, and it would be good.
[snip] if you examine WDM-PON’s composition, including its ability to support symmetrical transmission, then I believe you’ll change your views about the attributes you’ve highlighted and conclude that the limitations you cited are mainly a function of choice in implementation, and not those of inherent limitations.
COOK”s Edge: Complexities everywhere. Wouldn’t it be nice to have some open source flow charts available to allow non specialist planners to parse better the intricacies of the choice they must make?