If the single-channel transmission rates that you speak of were achievable over fiber using solitons, we'd still have a very long wait before the opto-electronics at the user interface were capable of catching up to them.
High throughput rates over longer distances has been the more appealing prospect of soliton transmission, not increased throughputs on a single channel, per se. Single-channel soliton technology will not aid in the creation of discreet paths for individual carriers and users purposes, though, which are critical to the underlying economics governing the sale of optical capacity over fiber strands.
Carriers selling at the wholesale level, and even those who provide their enterprise end users with lambdas at the retail level, for example, will still require some means of breaking down spectrum for individual user allocation purposes. In this respect I'm referring to dedicated optical channels. In the optical space this is still best done using WDM, aided, as time goes by, by associated technologies that can selectively filter and route lambdas on the fly.
Partitioning could also be done over a single channel, as in the instance you cite, providing, however, that some means of interleaving individual users is taken into account and provisioned. However, interleaving presents other problems, which are best documented in the well-known shortcomings of traditional forms of time division multiplexing (TDM) and statistical TDM.
In the end, the backbone provider still needs to groom multiple flows in order to provide their largest users their own lambda (lane) on the aggregate flow (multi-lane highway). There are many ways to achieve this, but single-channel soliton delivery at the rates you speak of will not be one of them until a means of routing user packets at those speeds can be done through optical means, and until such a means of delivery becomes palatable to users who demand that they be given their own "channel."
Likewise, interleaved TDM or STDM channels do not permit optical 'transparency,' which will be essential to a growing number of lightpath transmission services in the future.
Even so, the serial soliton link would still require deconstruction into its constituent user flow parts at some point (as it begins to leave the core), and those must then be parceled out to each user at the edge. Furthermore, the use of TDM and statistical methods (using available free time on a given channel) of sharing a serial (soliton) connection do not sit well with second- and third- tier carriers and large enterprises who want their own end-to-end, dedicated channels. Read: IRU (indefeasible right of use) lambdas.
The most sough-after benefit of soliton transmission remains its ability to support ultra-long haul transmission, spanning many thousands of kilometers, while obviating the usual need for regeneration at traditional mileage interval$. But here, too, other technologies are finding their way into the ULH space, as we type. Raman amplification is but one such alternative. |
