October 1997

Upcoming Battles Over Internet Access

BY BROUGH TURNER

Last month, I argued that IP telephony — the use of Internet technology for telephony applications — will supplant traditional telephony, revamping the telecommunications infrastructure in the process. The time frame for this transition? A decade or two. Yet the outstanding technical issues and standards work should (as I suggested) occupy us for only another three years. Why should IP telephony take any longer? The delay depends on how quickly we can get the raw bandwidth and the local connectivity we need to support IP telephony.

The Internet backbone is growing dramatically — both in capacity and in traffic handled. In North America in 1996, data bits exceeded telephone network voice traffic for the first time. So, backbone capacity will be there when needed for IP telephony. Local access, however, is another story.

With respect to local access, today’s big debate is over which technology is better: cable modems or some kind of digital subscriber line technology (xDSL). (See Table 1 for xDSL definitions.) The current debate neglects the long-term impact of wireless. Worse, as a technology debate, it misses the point. “Local access” is a political and regulatory issue. Local access is dominated by the politics of the current telephone companies, cable companies, wireless operators, and (to a lesser extent) other utility companies. The rest of this article focuses on the local access methods that will win out as feeders for tomorrow’s Internet.

CHOICES WE’D CONSIDER WERE WE TO IGNORE THE USUAL CONSTRAINTS
Decisions over how to implement local access aren’t strictly technical. As just mentioned, our choices can be constrained by politics and regulations. Another key issue is the existing infrastructure. This issue is particularly important in developed countries, where there is a lot of existing infrastructure — dialup connectivity, telephone loops, cable TV, broadcast TV, cellular phones, satellites, electric power wiring, and so forth. But let’s put all that aside for a moment and assume we could start from scratch. Let’s imagine what our choices would be were we to bring high-technology to an undeveloped country.

All communications networks get economies of scale by multiplexing the communications requirements of multiple users — that is, combining many users’ communications onto a shared infrastructure. The closer the multiplexing is to the users, the more money you can save when building the infrastructure. So, access methods with inherent multiplexing are more costeffective. One important advantage of the packet-based Internet is that it multiplexes traffic from many users more efficiently than the PSTN.

Another consideration is the ever increasing performance of electronics. As a result of this progress, wireless technologies are gradually overtaking wireline technologies. And, note that with wireless technology, multiplexing is inherent in the local access infrastructure.

If we could ignore political considerations and issues over existing infrastructure, the choice of access method would vary based on population density. In the long term, however, the infrastructure is likely to use wireless technology locally and fiber in the backbone. This trend is already in place for voice communications in developing countries. Internet access will not be far behind.

Satellite-Based Radio
For extremely low-density, widely spaced populations (for example, ships at sea), the most cost-effective solution is satellite-based radio. This approach is already in use for telephony at sea or in remote locations.

Land-Based Radio
As population density goes up — say, a rural farming area — the best solution shifts to land-based radio, using cellular principles. This makes far more efficient use of the available bandwidth than satellite-based radio, as a small block of radio bandwidth is used repeatedly, in adjacent communities. As the population density increases, more and more radios are added (more “cells” in the cellular jargon). But the power requirements of individual radios go down due to shorter distances (so the radios cost less). Thus, while there is more total equipment, the cost per subscriber goes down.

Fiber To The Building
At some point, it becomes more economical to run fiber directly to each apartment complex or office building and then use either wired or wireless LAN technology within the building. Using radio, infrared or other wireless technology for the last 30–100 feet will become more competitive as the cost of the electronics continues to drop.

OUR CHOICES IN THE LIGHT OF CONSTRAINTS
Of course, decisions over local access are not strictly technical. If we are to review the real options for a developed country such as the United States, we need to account for the existing infrastructure and the role of politics.

Dialup Access
Currently, dialup access via a modem is the most popular method for subscribers in their homes to get connected to the Internet. The modems are typically 28.8 Kbps modems, evolving perhaps to a maximum of 56 Kbps. IP Telephony is possible at these speeds, but when you add the hassle of making the dialup connection, it’s not a compelling service. A “permanent” connection at higher speeds would certainly help accelerate deployment of IP telephony.

The 28.8–56 Kbps dialup limits are set by the equipment used to terminate the twisted-pair copper wire from each home at the central office — equipment that is based on what was state-oftheart voice telephony in 1963. The copper wire itself could carry more bandwidth, but the rest of the PSTN infrastructure restricts the achievable bandwidth. Luckily, there are better bandwidth alternatives that will eliminate dialup access as the predominant way to access the Internet.

ISDN
Basic Rate ISDN (BRI) changes the termination at the central office and supports 144 Kbps over the copper wire. BRI was defined based on 1980s technology. In the late 80s, it began to make economic sense, and people tried to roll it out. There was an expectation that by the early 90s, ISDN would be everywhere. However, due to the regulatory and bureaucratic issues, that did not happen. Today, a few European countries have fairly broad ISDN deployment, but there is still virtually no deployment in the United States or elsewhere.

ISDN, in my own experience, has been less than inspiring. ISDN connectivity to the Internet from the home is fairly expensive (over $80 per month), certainly more expensive than other options. Also, service is sometimes unreliable. Basic rate ISDN is a technology whose time has passed — not the ultimate solution for Internet access.

xDSL
Currently all of the major telephone companies are looking at xDSL. xDSL uses advanced modem electronics at each end of an existing copper pair to increase its capacity from 56 Kbps up into the 1–8 Mbps range, depending on the condition and length of the copper wire. Globally, the telephone companies have a window of opportunity to reuse their wire, but whether or not this will happen is a political question. If xDSL deployment is left to the monopoly telephone providers, like the RBOCs in the United States, then it will happen just about as fast as ISDN happened, that is, hardly at all and in only a few countries. All of the problems of deploying ISDN technology apply to deploying xDSL technology. In addition, xDSL would cut into existing revenues. Who would pay $1,000–$2,000 per month for a T-1 line if they could get xDSL for $50 per month? So, don’t expect competitive xDSL from your local RBOC any time soon.

In the United States, the FCC is a wild card. They would like to leverage the Telecommunications Act of 1996 to give third parties access to existing copper facilities. In other words, third parties could lease bare copper wires from the existing RBOCs and connect xDSL or other equipment of their choosing. This is currently being fought in the courts because it would put a big dent in the telephone companies’ monopoly. If it became possible to lease a twisted-pair copper wire to a subscriber’s home for a reasonable price — say, no more than it currently costs for that copper pair with basic telephone service on it — then xDSL deployments will occur. Multimegabit downloads will be available at much less than $100 per month. But this is not a technology or business issue — it’s purely political.

Cable Modems
Another existing infrastructure in developed countries, particularly in the United States, is the coax cable used for cable TV. Cable TV networks were originally designed to be one-way only — broadcasting TV towards the home. Making a two-way infrastructure out of this requires that the cable manufacturers retrofit many components in their network. Then they need to place cable modems in each home that subscribes. That’s a significant investment, but it’s not as bad as laying new cable. In addition, the cable TV coax has more potential capacity than the twisted-pair cable used by telephone companies. It is also more efficient because data is multiplexed closer to the user. In cable TV, multiplexing happens within 100 feet of the house, while in a telephone network it could be miles away.

Cable modems providing 1.5 Mbps of bandwidth are being deployed in a half dozen major U.S. cities today, with Internet service included, for $40–50 per month. Note that this is incremental revenue for the cable companies, so there is a strong incentive to deploy the service! Personal experience and anecdotal evidence is that cable modem Internet service costs less and is more reliable than ISDN-based Internet service, at least in New England.

Can xDSL compete with cable modem service? Perhaps, but only if third-party economic interests gain access to the raw copper wires in the near future. Cable modem technology already has a cost advantage, and a two- to four-year lead, at least in our area. Cable modem Internet service is here today, while the RBOCs are still talking about trials and fighting thirdparty access to their copper.

CONCLUSIONS
Serious alternatives to dialup access are appearing today. In the next few years, we can expect major deploy ment of cable modems. xDSL will be deployed if third-parties gain access to the installed base of copper wires. If that happens, there will be a few years of exciting competition! Either way, we will see local access bandwidth and connectivity more than adequate for IP telephony.

Brough Turner is senior vice president of technology at Natural MicroSystems, a leading provider of hardware and software technologies for developers of high-value telecommunications solutions. For more information, call Natural MicroSystems at 508-650-1300 or visit the company’s Web site at www.nmss.com.