re: GbE used in Canarie, and other initiatives
Ray,
Yes, Gigabit Ethernet is being called upon to support new metro and other serving area topographies in support of a growing number of applications, such as distribution networks (cable tv, dsl-like, etc.), institutional networks aka I-nets, dwdm enterprise extensions in MANs, and so on. In addition to Canarie and a few other examples we've discussed here recently, not to mention the GbTTH-RM (reference mode) soon to be released by some participants here, see the story posted by Jim Kayne (thanks, Jim) this morning over in Last Mile:
totaltele.com
In each case the principal service providers and their vendors behind these undertakings are taking a somewhat unique approach, as they tweak the Gigabit Ethernet protocol in ways similar to how, in the past, Cable Labs prescribed doing so at the lower speeds of Ethernet, and in various other venues, to meet the needs of the applications they were striving to support.
At the same time, as witnessed by the Extreme story, one can easily say that the protocol itself is indirectly being tampered with by the means through which individual queues and device ports are parameterized and administered.
Then again, this is akin to what's taken place at lower Ethernet speeds in the past in the cable modem architecture where the rules of Ethernet were altered to meet the longer delay times to and from each customer location, which caused the elimination of certain "listening" and "backoff" parameters used in packet collision management, and in accordance with the dictates of the design idiosyncrasies of head end Cable Modem Termination Systems (CMTSs).
Of course, as this occurs we see a danger of fragmentation and a departure from the IEEE standards specifications, which could have the effect, if not orchestrated properly, of introducing chaos into the larger Ethernet model which in its original form is fairly simple and straightforward, which is one of the chief reasons for its widespread use, to begin with.
If standardization disappears or is compromised to a large degree, so, too, do the economies of scale go away that have been associated with its being easily commoditized across the board, as it has been to date in the traditional 10/100/and now 1000 Mb/s LAN spaces.
------------
The following statement from you did not go unnoticed, and is something that I must continue to remind myself about and constantly re-evaluate:
"All this city stuff is purely hypothetical, I want you to know, out here. We just had the irrigation water start to come. That's the big news out here."
It often occurs to me that my assumptions are the result of a sense of urban myopia that possesses me due to the influence of my NY City surroundings, and past experiences. And it's true that most of the star wars-like terabit material we discuss, as it relates to metro and end user applications, will generally be made available in greater volumes in areas where dark fiber carriers proliferate, and where some recently enlightened incumbents begin to see the light. Sometimes the incumbents will not do it themselves. They may rather prefer to partner with a Dark Fiber company, such as the case of BEL and MFNX. This model will be replicated all over the country, IMO, where it hasn't been either planned or implemented, already.
----------
On the surface, therefore, the perception is that rural and suburban areas would be among the last ones to see ultra high bandwidth being delivered to their doorsteps. And yet, we read stories like Canarie, and more recently how power company consortia are preparing to undertake builds in their regions with the hope of supplanting incumbent cable operators, at some point. I have a theory, albeit partially half-baked, about how fiber to the home will unfold. It states that dense-urban residents, especially those who are closest to telephone company wire centers and central offices, will be among the last ones to enjoy pure fiber to their doorsteps. Conversely, residents who will see it first will be those who reside in the 75 to 85 teledensity/proximity percentile, or those who are just short of being demographically too far, to ever justify it at all.
I say this because those who live across the street from the central office use twisted pair that will probably support vdsl at 52 Mb/s right now. I wouldn't look to the ILECs to improve on that capability anytime soon. In contrast, those who live in suburbia and deeper out into the boonies, are already seeing fiber being brought deeper into the field and into their neighborhoods, in the way of HFCs and Integrated Fiber in the Loop (IFITL) technologies using passive optical networks (PONs). It's called long term fiber creep.
We see this, for example, in BLS territory right now, where the next logical step would be to bring fiber directly to the curb, and then for those who can afford it at first, directly into the home (as BLS is already doing now in other deployments, as well).
So, as much as I am guilty of myopia when it comes to treating dark fiber in dense urban and metro settings, these other developments should also be kept in mind where fiber to the residence is concerned, since they demonstrate an inverse kind of relationship to that of dark fiber availability in the big towns.
------------
As an aside, a long time friend and business associate of mine, who is probably closer to large fiber sales nationwide than anyone else I know of, tells me that LU has stopped taking orders for new fiber from new, would-be customers because of the backlogs they are currently experiencing. Corning is equally swamped right now, as are the international suppliers. Orders for substantial size amounts of single mode are now looking at one year intervals and longer, even for most favored customers, he tells me. New FLECs are scrambling, as a result. You can draw your own inferences from this. It says to me that the new plants that Corning and LU have announced could have been implemented a couple of years ago, and they'd probably still be working at capacity at this time. I'll probably be voicing the same retrospective in another three years. And then another, as new parameters having to do with permissiveness, or nonlinear handling attributes, or whatever, are perfected.
I guess these silica barons were just helping to thwart off a premature fiber glut, eh?
FAC |
