Andy, assume for a moment that I was able to order and obtain an all-optical network from coast to coast for my intranet. With MFNX and Williams, and now several cable companies and utilities getting into the act, how many different grades of fiber do you suppose I'd find on an end to end connection between NY City and San Francisco, demarc to demarc? How would you characterize that sort of connection? Would any single fiber type stand a chance to be present through and through? Or, would it be more likely that in NY I might find myself launching on SM-28 in the access portion, being picked up by LEAF on the first long haul link, then met by ALLWave on the second long haul link, and then onto MetroCor in San Francisco?
I think for your argument to hold up (assuming that it had merit, and I've not looked at it that closely) it would require that you had a continuous set of tandem connections all being of the same type of fiber. Reality is somewhat different. For example, here, in NY City, we have three different grades of fiber being offered by the same dark fiber carrier, with each grade tracable to a point in time when that particular grade was known to be the "state of the art."
While these conditions are not uncommon and have been tenable during the days of SONET-only networking, they will begin to impact on more enhanced forms of optical networking in the future because of the discontinuites and mismatches that will ensue -- in terms of how one grade handles dispersion/PMD/etc., and how the next one in the chain that it is connected to, does. Herein may actually lie some additional reasons for some vendors' wares not working very well, as well. Or, these conditions may become the prime cause for adhering to "selective engineering" practices going forward, when choosing strand facilities for links belonging to a specific optical network.
FAC |
