Skip navigation

Breaching the digital bottleneck

Imagine processing 400-billion pieces of information per second! Sound impossible? That's exactly how fast backbones used by today's Internet providers operate-at speeds of up to 400 gigabits of data per second. What's more, in a year or two, they'll be even faster. However, this state-of-the art technology does not exist everywhere. In the real world, our high-tech offices appear to "blaze" at a

Imagine processing 400-billion pieces of information per second! Sound impossible? That's exactly how fast backbones used by today's Internet providers operate-at speeds of up to 400 gigabits of data per second. What's more, in a year or two, they'll be even faster. However, this state-of-the art technology does not exist everywhere. In the real world, our high-tech offices appear to "blaze" at a tortoise-paced 1.5 megabits per second, and the last link to the home struggles along at a snail-paced 30 or 40 thousand bits per second. So while the computer gurus of Silicon Valley want to provide many things to us, this "last-mile" problem continues to tie them down.

This situation continues because the government protects local telephone companies (Telcos) from competition. Since there's no reason to change, they continue to make big money leasing 1.54-megabit lines to businesses. (Called T1 lines, these can cost from several hundred dollars to more than $1000 per month). In addition to this lucrative leasing business, Telcos also profit from multiple phone lines in residences.

If the Telcos allowed prices for bandwidth (the amount of signal sent from one place to another) to keep up with technology, they'd lose a ton of revenue. Remember, you and I cannot go into the telephone business. If we tried to open up a second (and better) telephone company, the government would shut down our company, confiscate our assets, and possibly send us to jail! Since competition is forbidden, normal market forces are not at work here.

The coming "end-run" years. Because of this Telco bottleneck as well as the inherent inefficiency of the telephone network (designed for the world of 1930, not the 21st century), the new world wide Internet data network simply cannot push through the morass of local telephone company systems. Instead, it's leapfrogging the telephone systems any way it can. This occurs in the following ways:

Through-cable television systems. This is a very good method from a technological standpoint; cable TV is a large pre-existing system with the proper type of architecture. Nonetheless, it's still pretty questionable whether the cable TV system will provide a good last-mile link. There are two main problems: (1) Because the government regulates cable TV, Telcos will probably donate enough money to politicians to make sure restrictions on the cable companies never go away. (2) Cable TV companies themselves have been rather predatory, generating significant ill will among customers. In fact, many people would rather not do business with them at all (if they had a choice).

Via satellite. Right now, there are several high-speed data satellite systems under construction that can deliver data to a home or office at rates well into the megabit per second range (perhaps even into the gigabit range). However, these systems will probably not be fully operational until 2002. After that time, they should effectively make an end-run around the Telco system. Currently, it appears they have not been restricted by the government. Therefore, there's probably not enough time left for Congress to shut them out of the market.

DSL. This is a telephone technology (as opposed to a data technology) called Digital Subscriber Line. It uses electronic boxes on each end of a standard telephone line to achieve speeds of up to 1.5 megabits per second. This technology comes in several flavors (ADSL, HDSL, etc.); however, they all are somewhat similar. Whether this will ever by implemented is questionable. It's really not in their best interest to do so, and Telcos might not readily embrace the technology. For example, Telcos left ISDN, the predecessor of DSL, to die on the vine. The same fate may await DSL. We'll just have to wait and see. Digital power line (DPL). This new technology uses a spread-spectrum power-line carrier technology to send up to 1.5 megabits (and in some cases more) over standard power lines. This system will probably be too limited for use 10 years from now, but it could be a major factor for the next five to eight years. As far as this author knows, there is no legislation limiting the use of this technology. Besides, it operates over existing power wiring, which is an enormous advantage.

How the DPL works. To send over a million bits per second through power lines, DPL uses what is termed a high-frequency conditioned power network (HFCPN). Rather than trying to push one signal very fast through the power lines, it breaks the signal into multiple streams, each of which is sent through the power system on a different frequency. Special receivers combine the mini-streams of data back into the original signal configuration.

These networks are rather complex but extremely stable and tolerant of electrical noise. They can provide between 6 MHz and 10 MHz of usable spectrum to far-end customers and more than 20 MHz of usable spectrum to near-end customers, for peak signal power levels of between 1mW and 10 mW. The overall spectral efficiency of the network depends upon a number of things, including:

* Customer type and number per distribution unit (typically up to 50).

* Type of multiple access requirements (dedicated or switched).

* Service requirement (voice, data, still or moving pictures, etc.).

* Digital or analog transmission technology.

* Modulation, coding, and compression schemes (data bits per unit of available spectrum).

* Traffic density mean and peak.

Under this scheme, you terminate the network interconnection conductors in a three-port directional coupler, known as a conditioning unit (CU) (Fig. 1). The basic elements of the CU are shown in Fig. 2 and include both high- and low-pass filter sections interconnected to form a frequency sensitive directional coupler, which has a network port (NP), communication distribution port (CDP), and electricity distribution port (EDP)

These CUs provide for the following:

* Safe and efficient interconnection of signals at speeds greater than 1 MHz.

* Directional propagation of signals.

* A reduced noise floor above 1 MHz.

* Isolation of variable customer loads.

* Suitable network service termination points for electricity and telecommunication services.

* Optimum spectral performance of the cable network.

Implementation. To implement DPL, utilities will require an investment in equipment and training. However, this investment will allow utilities to get into the Internet provider market more quickly.

The number of electrical utilities who will want to jump on this bandwagon is questionable, but you can certainly make plenty of money in these markets. Because the electrical utility market in the United States is deregulating for the first time, there are plenty of changes on the horizon. Most notably, competition will enter the market for the first time. As power markets become competitive, taking advantage of new opportunities will become important, and utilities will take a careful look at technologies like DPL. Hopefully, they will implement them quickly.

The possible utility company end-run. Earlier, we discussed methods of making end-runs around the last-mile data bottleneck. If they choose to do so, electric utility companies could provide a terrific way to get high-speed data to the house or office. They already have cables entering every home and office, and they have plenty of expertise running cables over long distances.

There are a lot of reasons why it makes sense for electric utilities to deliver data to their customers. The tree-and-branch architecture of power distribution systems is perfect for Internet data transmission. In addition, it's now very economical to run optical fiber along long-distance transmission lines. Special optical cables are available that are either combined with transmission ground wires (the top wire on long-distance power lines), or wrap around the ground wire. Many utility companies have begun doing this in the past several years, and some have a lot of fiber already in place. Initially done for communication between substations, other users now rent out some of the fibers. If you install enough fibers, it could form a very significant communications structure. In fact, it's quite possible electric utility companies could challenge Telcos for all forms of communications business in just a few years-and win.

Telco technology is not an efficient way of delivering digital communications; it's built around large, expensive central switches. Remember, networks are routed rather than switched. In other words, each piece of data contains its own destination address and routes toward its final destination at several points during its travels. The Telco network is slow, expensive, and smart. The routed network is fast, cheap, and stupid. The necessary intelligence is not in the network; it's attached to each end of the network. Fast and cheap is better than slow and expensive. And distributed intelligence is better than centralized intelligence.

What does the future hold? How all of this actually plays out over the next decade will be fascinating. At some point, political demagoguery will inevitably come into play: "The communications infrastructure is a national asset-we can't allow it to be turned over to people who are interested only in profits," or "Preserve it for the children," etc. There will probably even be several personal and financial battles. The forces of regulation and control cannot hold out indefinitely. But, they won't go without a fight. So, stay tuned.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.