Dan, just so we don't lose sight of the origins of this discussion, it did start with the news of ATT and TCI, right? In any event, I believe we're talking about the economics of cable telephony and IP voice. The same arguments can also hold true and be applied to DSL and wireless.
You've sized it up pretty accurately from an historic perspective, but the rationale used for DS0 equivalency pricing may be seeing its last days, where such comparisons as the following are concerned:
>> However, the point I am trying to make is that pricing on circuit switched networks is usually based on as DS-0 equivalent. Because a DS-0 equivalent (i.e. 64kbps) is only 1/100th of the data rate needed for video, then video transport becomes expensive if pricing is based on the DS-0 equivalent circuit price.<<
In order to reach a price or "rate" per DS0, or a 64 kbps voice slice in the past, there have been "rate element" contortionists at work, steadily, for the last eighty years or so, whose jobs have been to compose and submit tariffs to the regulatory agencies. There was no other game in town, so their say-so was gospel when it came to costing out the elements that made up the individual parts of the whole.
Today's competition, fostered and in combination with the economies that are being achieved through technology, are collectively beginning to erode those rules of the trade, and newer [market] forces are beginning to take form.
Also, due to the changing proportions of voice to data, with data taking over the dominant role in the greater network, and the reduced b/w requirements for the newer voice services, we can expect to see a continuing shift to lower and lower costs for voice in comparison to the greater bill that will be paid for data and video. That will hold true for businesses as well as residential. In some instances, voice will indeed be given away for free, or more likely will be used as a wild card somehow in the negotiating game.
Based on the way that VoIP works, in fact, it is cheaper to provide than equivalent amounts of "bits" and their associated bandwidth used for data, since VoIP uses User Datagram Protocol, or UDP, for the most part, and most other forms of IP Data use Transmission Control Protocol, or TCP. The analogy I like to use is that UDP packets are able to get through the maze like cockroaches can get through cracks in a wall. TCP, on the other hand, looks for guarantees... it sends out scouts to survey the uncharted terrain, gets feedback and acknowledgements, etc., and takes more time and bandwidth in the process.
Not an elegant means of explaining the differences, by any means, but effective. In effect, UDP packets can squeeze through tight spots where TCP cant, because TCP packets must window and undergo slow starts, etc., making them undesirable, and more costly (both from the networking dynamics perspective and the hard currency perspective) than UDP. But TCP has been absolutely essential, up until now at least, for passing reliable other forms of "data."
To net it out, UDP is cheaper, albeit less reliable, packet by packet, but some packets can get lost in voice streams without noticeable effects, and overall, UDP has been demonstrated to be the more appropriate form of transport for voice where unpredictable Internet Protocol environments are concerned.
>> If they are both carried on the same transport, switched by the same routers (Not TDM circuit switch) then how is voice going to be priced and how is video going to be priced? Sprint says that it will be priced by the bit transported or the data packet transported (somewhat like X.25).<<
Routers are not the democratic animals we may take them to be. Policy based decisions for alternate forms of payloads differ with respect to costs (again, of all types), and those which require isochronous-like or real-time or n-rt deliveries, will be more expensive than those which are more casual, and those which use UDP. This is the focus of much discussion and debate, and more than just a little consternation inside the decision offices of ISPs, concerning tiering of services, and the assignments of premiums for higher grades of services to subscribers of Internet access.
>>Either voice calls are going to be dirt cheap or video calls are going to be very expensive.<<
Right off the bat, today we are planning on Voice over the Internet requiring only one tenth the bandwidth that it did under the DS0 Cage rule. That knocks out one assumption right there, so naturally, it will be cheaper to acquire than it did before without any magic taking place.
It could be further reduced in cost if it is assigned to a channel transport on the basis of syllabic detection, like speech activated cut-through (no bandwidth is assigned during quiet periods, or greater than 65% of the time; do the math, if you like), and then its efficiencies are actually further doubled, or further quadruped, or more, on a statistical basis, depending on what the aggregate capacity of the pipe is; and how many other talkers are in contention; what the speech detection criteria are; what the overall constants such as the overall latency of the path is; etc.
Likewise, certain video compression techniques are only using bandwidth when there is activity to convey, and even here, the algorithms are set to only send information content "change" information about the picture, and not the entire picture, frame by frame. Obviously, you are correct in that video will be more costly, but the actual formula to determine what the differences are will be largely market driven criteria, looking to competition and the technological advantages that can be attained, rather than the reference to established rate element information in the offices of the state PUC.
>>Carriers that have big voice revenues must be worried about how they recover from the apparent big decrease in what used to be DS-0 equivalent voice calls.<<
Amen. I couldn't help think about this today when I listened to the ATT/TCI conference call. These folks are espousing full blown VoIP over Coax Cable and Wireless Local Loops sometime in the next two years. Ubiquitously! The only surge of switched services they say will be where TCI has already created an excess capacity of switching from their previous foray into voice. Oh, and in the deepest reaches of rural coverage, where it will be uneconomical to provide the two-way and necessary upstream capabilities.
Where on Earth are they going to find the means to restrict their subscribers from gaining access to other ITSP's capabilities, once they are already on the Internet through high speed STB/Cable Modems and air interface devices? Once you let the cat out of the bag, it's going to be difficult to get it back in, or in this instance, for them to contain or control their own subscribers' choice of LD (and other service) providers, including other means of _local_exchange_ provisions over VoIP!!
The answer to this churning of accounts, they believe, will stem from single billing conveniences through bundling. Huh! And at the same time, T is now providing prepaid VoIP services in a growing number of markets which they say will cut down on back office expenses for them, with all the conveniences for the subscriber that could not even be matched by having a bill sent home in the first place <?> !
[All, you can still get to hear this recorded conference call at 1-800-475-6701 pin# 396884 until Sunday Midnight]
>>I mean equivalent cost per bit will have to be reduced by 95+ percent if video and other high band width consumer and e-business applications are affordable. Right?? <<
Sometimes. All bits are not alike, though, as I stated above. Sometimes a whole bucket of spent bits aren't even worth two bits. <smile>
But your points are well taken, and many of the rules of thumb you've cited will endure only as long as the lack of competition will permit. Once the new players have their way at competitive pricing, there will be no looking back. We can see this happening already where VoIP is allowed to take place.
Regards, Frank Coluccio |
