WTC,
The reason for my seemingly schizophrenic approach to both the wire-line build-out model by DLECs, and the emergence of a virtual unbundling paradigm by the upper layer-based service providers, is because I believe we will see both of these delivery forms evolving in their own generalized ways, in the near future. Probably in symbiotic fashion. Each of these is demanding of their own disciplines, since they are not alike at all. Rather, they build on one another.
In areas where independent telcos prevail, the smaller mom and pops that is, this may even be seen as a form of welcome relief wherein the smaller ITOs may find it beneficial to leverage in a form of reverse unbundling. Hey, it's happened before in other networked contexts.
I'm certain that the "fourth buildout" will not be a monolithic one, though, as I've suggested before when I was first presenting the proposition, for that would require unreasonable levels of cross-industry and government cooperation that would probably in all likelihood become politically unviable. But in a collective mode, wherein several different generic initiatives are launched in diverse regions, the equivalent effect may in fact ensue in some locales. Witness Sprint's most recent statements that they would target their own 19 serving areas first, and target the foreign markets later.
And there are many independent utility company and municipality buildouts taking place at this time too, to take up some of the slack.
If one wants to reach a bit, there are some precedents for cooperatives of this nature in other networking venues. Several that come to mind, where large numbers of smaller players were involved, were:
(1) the National Telecommunications Network, or NTN, an alliance of about fourteen regional fiber optic carriers primarily in the Midwest who, during the Eighties, banded together to unify their lobbying powers and joined their networks together to form a coherent, and interworking, platform;
(2) earlier this decade when multiple, smaller IECs and VANs of varying sizes formed a Frame Relay consortium called UniSpan; and,
(3) 'lo and behold, the Internet itself.
Neither of the first two mentioned above have survived the test of time, that I am aware of. Maybe someone with more knowledge than I can elaborate as to what ever took place there, in their disappearances. I can't find anything on them except a web page for Unispan that is still under construction.
But the Internet is getting larger with each passing day. [Curtis, don't even think about it. ;-) ]
Back on point, I don't think that this is a stuation that must be limited to only binary states, where only 0 and 1 are allowed, where 0 = a fourth independent buildout, and 1 = virtual unbundling.
The other schemes which prevail today will continue to thrive, including modems, V.90, ISDN, wireless and to a more limted extent, satellite, although they will not be "logical" solutions which can be accessed as "services," instead, they will, for the most part, be used as roadbed.
I'm referring to ILEC deployments of DSL and the CableCos expansion of their HFC/CableModem capabilities. And wireless local loops and some forms of satellite delivery will also grow in share, but not as significantly as the first two as far as I can see at this time... or will they?
To make the distinction from what I propose throughout the rest of this post, however, the former are physical layer disciplines which need not be tied inextricably, or in any way, necessarily, to the delivery of upper layer services by the same provider. And here, upper layer services can be thought of as voice services, web access, fax, content, etc.
The compelling thing to keep in mind here is that each form of facilities-based delivery will serve as the lower layer foundation upon which, inevitably, virtual unbundling will take place, where discrete user services are concerned, once the end user has suitable access to the 'net. Or once they have suitable access to the newly (at some point) hybridized HIPSTN, or Hybrid Internet-PSTN.
The local access, in turn, during the next gen needn't be anything fancy or ultra high speed. It simply needs to be from a facilities-based provider supplying one of the previously mentioned media forms at super 128 speeds. If this is satisfied, 'virtual unbundling' will very likely come to pass whether the dominant facilities providers endorse it or not, if not from specialized ISPs, then from the regular ISP of choice, itself.
How long do viewers here think it will be before AOL itself begins to offer VoIP services to its members? Just another brick in the retainer wall at some point, I think. For if they don't, someone else will.
Thus far, they have only entered into a co-marketing arrangement with a traditional switched voice service provider.The same provider, or someone else, could very likely become the agent of change whereby VoIP is offered by AOL to its end users in the future, in another form of co-marketing agreement. Or maybe AOL will take it in-house, themselves.
There are many provisos here, the most obvious being that the end user must have ample fundamental, and relatively unobstructed, bandwidth in both directions (it needn't be super-fast, or even symmetrical, however) to be supportive of real time voice, data, low-speed video, etc., applications, and they must be able to support sizable transfers in each direction.
High-quality multimedia will prove to be a bear for any platform in the future, unless fiber is pushed a lot closer to the residence enabling higher DSL [vdsl] speeds, or, on the other front, unless the cablecos do some serious re-segmenting of their outside plant to eliminate or drastically reduce the contention problems they will face in the coming months and years. Wireless speeds, too, will need to be higher than most WLL providers of any promise are now touting, while at the same time remaining economical to the user.
Therefore, I think that Internet access and voice services will continue to be the principal areas of short term interest here.
Stated another way, Internet access alone, via suitable intermediate speed access lines, could soon be sufficient to allow end users to begin picking their voice and data providers from whomever they choose without dependency on the ILEC or CableCo. Likewise, this requires that said discrete service providers begin to make these services available, soon, as well.
[[While we're on the subject, I do believe that DSLs are better suited at this time to deliver web access with consistent quality, even if they do not peak to the highest idle-period potentials of CableModems.
This holds true for access to the traditional forms of data and voice [in the case of split function ADSL], but DSLs, too, will pale in comparison to the demands that will materialize when it comes to serious multi-application access requirements in the future, especially when simultaneous forms of video + other multimedia apps begin to appear.
We'll most likely see a form of leap-frogging take place by cablecos when they get it together with a massive re-engineering effort, only to prod the ILECs to yet another push of fiber closer to the end points. It's a hell of a migration path, wouldn't you say? ]]
Of course, all of this assumes that the facilities-based providers who'll be supporting all of these capabilities at Layers One and Two in the distribution plant will actually scale their central-office and head-end nodes accordingly, and make similar provisions in their upstream access connections.
I.e., they will connect these newly re-sized nodes to an even greater abundance of bandwidth in the surrounding network "edge" networks, and to the core, to be able to handle all of what's being said here. It is at this point that another face-off will occur in the information handling supply-chain that we have come to know as the Internet.
There will be structurally separated areas for access providers, discrete service level suppliers, ISPs, backbone network providers, etc. This will force a showdown which will concern, where it hasn't already, whom, exactly, will bear the burden of supplying [read: engineering, implementing and ultimately paying for] that capacity between the virtual providers [who may reside deeper into the cloud or cross-country] and the facilities-based providers who must maintain the outside plant, and in some cases the wireless facilities.
Assuming that all of these issues are handled in an equitable fashion, and once we get proportionately beyond the V.90 56k model, the selection of service providers at the upper layers becomes feasible.
I may want to use, say, the Bay Ridge voice gateway service to long distance, instead of Bell Atlantic's Class 5 switch to my presubscribed Interexchange Carrier. One of the appliances hanging off my PC, connected to my residential premises wiring system, would make this possible for me in a transparent way... using one of my wall phones, or I could do it right from my desktop.
Or, my distributed so-ho organization based on a virtual office concept may elect to use a virtual call center service from our favorite specialized ITSP, rather than an 800 redirect service from an Interexchange.
In this last instance, my work force may be comprised of work-at-home part-time agents who are all answering incomings on their desktops, or from a console of a specified nature... just as they do with regular 800 virtual call center applications which employ traditional AIN, SS7 and POTS lines.
The ITSP model would be independent of the Class 5 functions of the ILEC, whereas IEC facilities access would almost demand a certain level of end office functionality be involved. Granted, however, not necessarily, though.
But not so when the ITSP is used, with the possible exceptions of dependency, still, on subscriber line data base lookups and SS7 access, made possible through some still-to-be-formalized directory services integration between IP and PSTN identifiers and pointer functions.
Some users of DSL and CM will undoubtedly be better off with a symmetric scheme when using these lines for telecommuting or when extensive video for personal use becomes a factor. In fact, they will demand it. But these tend to be rather relative assessments, in the end, because the bar continues to be raised.
What was once considered to be good as a symmetric upstream speed, may turn out to be lower in two years than the lower speed side of an ADSL line or future CM links. But these would be the purview of the facilities-based providers.
Up to this point, I think I've demonstrated why it may actually be necessary to use, like you say, "a belt AND suspenders approach in order to worship the god of Competition."
They are not mutually exclusive, since they operate at different layers of the same solution... which, in effect, actually make them co-dependent to one another, even symbiotic.
IMO, one of these can go nowhere, in the not-too-distant future, without the other. That is, the virtual unbundling camp, even though they do what they know how to do best, is hopelessly lost without a viable form of physical layer foundation made available by the incumbents (or others who are fool enough to try to compete with the incumbents on their own terms).
And the physical layer providers, "to be true to themselves in the end," must focus on what they, too, know best. Let me digress on this point a bit.
I've seen questions raised on other boards here in SI and elsewhere as to why CSCO doesn't take on the fiber barons on the latter's own turf. And why, say, LVLT doesn't go all out and do the same thing that QWST is doing 'at this time,' that is, shot-gunning the Earth with silica strands, instead of perfecting a Layer 3 networking model for others to follow.
And why LU doesn't just go out and buy, outright, some of the carriers who they've been financing (and are not doing so well, presently) instead of simply watching the outcomes from the side lines.
Also, I was recently asked why MFNX doesn't engage in a fully electrified level of SONET-ized and Layer 3 activities, instead of "just pulling fiber." (Actually, MFNX is active in WDWM, including metro situations, but that's where they stop with their basic frame work of services for the moment, preferring to offer strands and lambdas as a carrier's carrier, instead of becoming one of those of a class to whom they are currently selling.)
I think that the answers to all of these questions are clear. The principal answer, of course, is that these players are stars in their own rights, because they have focused on their core missions, no trivial sets of tasks themselves, and to a large degree because they did NOT venture into unchartered and unfamiliar territories that were already the cultivated domains of others. And the other factor is that to do so would, in many cases, be violating a rule that states that you never enter into a field that relegates you a major competitor to your best customers.
This argument extends into the present discussion as well. It is my belief that no one knows the outside distribution plant better than the telcos, and no one knows broadcast and cable TV delivery better than the MSOs (okay, with this one I went out on a limb a bit where some providers are concerned, as I did with some of the ILECs, but hand in there with me).
But when we begin to discuss last mile disciplines and emerging next-gen interactive platforms that extend into the abyss of cloud structures, that's where my praise for Telcos and CableCos begins and ends.
These companies are the best at what they do in the last mile, perhaps, but they are not the best candidates to deliver interactive services on a global scale, where logical functionality in the Internet IP model replaces those of time-slot interchanges in central office platforms such as DCSs, Class 5s, and other miscellaneous gear found in relay racks.
On the surface, and perhaps there is good reason for this, there appears that there could be no greater antithetical form of mismatch, than to impose the chores needed to be performed in the next gen environment, on the typical BOC organization, without some almost impossible changes taking place first. Not by a very long shot, and not without severe measures that would involve major cultural and philosophical surgery, which, historically, these players have never even acknowledged as possibilities, much less ever entertained. And I would go a step further in suggesting that perhaps the public has very likely benefited from this, in some perverse way, since these same organizations have been able to grow and maintain the most robust basic service infrastructures of their size ever known.
---
I'm not surprised that you would be left daunted by my suggesting that "a new carrier or consortium of carriers can economically build last mile plant and profitably serve residential telecom service demand."
In retrospect, I see that my wording belied my intent, as I most likely meant it at the time in a metaphorical sense, if not an actual one, but making the point took me to extremes... it happens sometimes. But no harm done, because it underscored the principle.
>>I say residential, because I contend that the business world is already well served in many urban markets by a variety of Competitive Access Providers, wireline and some wireless,<<
Again, these are the physicals. They may persist for a long time to come, but they cannot dictate to the end user which logical selection of services they may subscribe to, once the next-gen'ers are out there in force, and once the end users have adequate access lines in place.
---
You have rightfully taken my propositions at the word level, and I concede that some of what I stated previously was ultra hypothetical. Granted. For example,
>>I'm just a bit unclear where the new competitor is actually building new plant. Is the new competitor (let's call them a CLEC) securing access to ILEC (and possibly MSO??) plant at points convenient to their business plan? <<
I would suggest that there are a preponderance of startups toady, and who are now getting past the calculator what if stages and who will be filing for incorporation during the next six months, who would follow this very tact. They would attempt to emulate the models from whence they came (many are ex RBOC and CLEC folks). These are the ex-telco- type craft and management (like you and me, maybe some more enlightened, maybe some less [g] ) who come from the environs where this makes the most sense. That is, to do the same thing that they have been used to doing for the past umpteen years.
These are not the service providers who the ILECs should be overly worried about, however, for they can never beat the ILECs at their own game, when viewed from an economies of scale perspective.
The ones who should be viewed as most threatening to the ILECs (or who may actually allow them the greatest opportunities, depending on how the ILEC treats this situation), on the other hand, are those who DO NOT build out their own wireline distribution plant, or who DO NOT depend on physical points of presence, but who, instead, will leverage off the end user's already-existing connections to the cloud for the delivery of upper layer services and content.
At the risk of being repetitive re this matter, this separation of physical and logical is the major point of distinction I've been making all along.
Nonetheless, and to keep the argument going, there will indeed be those who continue to contend for rights of way, as well as those who will insist on connecting at the UNEs, no matter where they are, no matter what the costs, and as long as they can attach their DSLAMs and routers at some point in the incumbent's locations or field distribution plant, that is less expensive to them than if they had to backhaul the entire assembly to their own offices.
>> Do you think the problems with "recalcitrant ILECs" are so beyond the reach of regulators that a new structure is needed, even at enormous cost? <<
I see that I actually used that word. he he. I meant, of course, calcitrant.
---
Take a look at what has happened over in the ATHM space.They've acquired Narrative. I think that it portends somewhat of a directional shift for the future that would be hard to ignore... that of the evolving interface at the customer location (integrated digital video) and integrated processing deeper into the cloud structure (server level data integration and interchange), and what demands this will make on the overall model going forward.
news.com
Up until very recently, we've witnessed multiple generations of ever-faster Intel-based boxes becoming one of the cornerstones of the Internet model, driving most of the parameters we've been talking about. Two of the basic factors in this Internet model, once we get past HTML and http, have been the desktop PC's internal processing speeds, and the speeds of access lines which are afforded to us by last mile service providers.
Another basic factor in this model which strikes us in the headlines every day, often multiple times per day, has been the Internet backbone itself, and its ability to withstand enormous growth pressures without succumbing to a Metcalfe-ian- like overload.
---
>>I think privately run and managed collocation sites are certainly a possible outcome... That could be a far better, i.e., more fair and transparent (but not necessarily cheaper to CLECs), solution to remote collocation issues.<<
Reality does strike home occasionally. Startups DLECs tend to think that they have an inalienable-like right to using Telco central offices, and they have this notion that it should be almost as free to them as they "think" the 'net flows have been up until this point in time for users. They should ask some of the Tier 2 and Tier 3 backbone provides and ISPs how free it's been for them, much less the smaller ones who are struggling at this time to meet the price caused by burgeoning bandwidth demands.
I don't subscribe to either point of view. In fact, there are several things in this regard that could really set me off, but I'll stay on topic here with regard to architectural issues. Suffice it to say: Of course, it will cost significantly, at times, to colocate in any worthwhile and secure enclosure.
>>Of course we have some today for Enhanced Services equipment that is not currently allowed in ILEC collocation space, so there is a model.<<
We may as well look at that model, since it is probably closer to the one that I am espousing than the one it will replace. Such ESP locations are top heavy with data processing gear, servers and mainframes, and they do not fit the traditional personality of a switching office. So does the next gen ITSP or VoIP operator's, although you are more likely to see unix workstations there, instead of IBM Enterprise Main Frames. Now, there's a thought...
My reason for drawing the analogy to spectrum upstream a ways back, comparing it to pole space, was because of the similar scarcity of supply of each.
---
>>You make an interesting point about qualifications. Qualifications and competency are certainly essential for running any part of a globally interconnected network. Network protection today seems based on certain gateway software and procedural safeguards (limited at best, and subject to some newsworthy failures) and reliance on the competency of peering network operators. There are also limits as to what connections IP nets can now make into SS7 nets, but that is obviously all changing,<<
Yes, it is all changing now. The qualifications issue will not become a sore point with users, especially those on the receiving side of calls, until the renegade operators who cut to the bone on both cost and expertise make their mark. We haven't seen this breed yet, only a few of them have migrated from the intranet stages with their wares [where it is much easier to preserve quality of service and overall performance], but they are in the wings, impatiently awaiting their cues.
Some packet voice providers today are providing excellent services, especially on the international side... some of them don't even promote themselves as such, preferring to sell generic services based on plain vanilla PSTN, Vo IP at times when dedicated IP backbones are in place, some are using proprietary ATM algorithms from vendors such as Nortel, and some are using other forms of low-bit-rate-voice of more traditional stock.
And then...there are those who are attempting to meld their public Internet IP voice services with those of the PSTN. The state of the art is not ready for this yet, and CoS/QoS hooks still don't exist to make this viable, if indeed it will ever be viable on the public Internet. But there has been nothing done to stop these folks from polluting the PSTN in this fashion, and based on the way public sentiment is going, there probably won't be. The market will dictate the outcome of this particular type of QoS dilemma, not the IETF or the ITU. I don't know how this will play out in the short term, though.
I'm running out of space on this page, and perhaps I've been running on too long anyway. I'll get back to you later with anything I may have left out. Until then,
Best Regards, Frank Coluccio |
