The Last Mile Problem: Current and Future Solutions
By Robert S. Leventhal, Ph.D.
President, Blueridge Internet Technologies, Inc.
Charlottesville, Virginia

The "Last Mile Problem" is one of the most bedeviling issues facing the success of both distance learning and the Internet itself. In short, the issue is how to get the state-of-the-art of Internet technology to the average American household, such that using the Internet is a productive and non-frustrating enterprise.

For the Internet to successfully realize the vision of a new medium for both commerce and asynchronous learning, the average household link must allow rapid downloads of extraordinary amounts of information, rapid processing and presentation of that information, and rapid uploading of responses to the information. While the 'gee whiz' element of the Internet still somewhat offsets the often painfully slow process of getting information on screen (particularly graphic information), the current level of frustration with this process is totally inappropriate for the medium's proposed future.

The high-speed links that characterize the connections between major nodes on the Internet, and the speedy hardware and software that processes information at that level have not exercised their muscle for the home user yet. The average household connection to the Internet consists in many cases of a relatively slow older PC connected with a modem allowing data transfer speeds of only 28.8 KiloBytes per second (KBps).

For the great promise of the Internet as facilitator of distance learning for the masses to be fulfilled, the speed of this end of the line connection (The Last Mile) must improve. Fortunately, there is good news on this front. The time is coming when the average household user will send and receive text, audio, and even video information with little to no delay.

The Adequacy of Home Computers

The current crop of PCs, whether the Intel chip-based Windows machines or the Motorola G-3 chip-based Macintoshes, running at 250 Megahertz speeds or more, is fully capable of processing all the information that a multi-media file contains at speeds that will satisfy home users. The bottleneck is not then the home computer, at least if it is an up-to-date version of the technology. It is primarily the connection to the house and the modem connection to the computer.

ISDN...A Well Publicized Solution

Only a year ago or so, the problem of "The Last Mile" was still being answered from many corners with calls for widespread application of Integrated Services Digital Network (ISDN) services. Using the same copper phone lines as standard telephone service, ISDN promised the delivery of up to five-times the speed of the conventional 28.8KBps modem, providing a raw data transfer rate of 144 KBps. In many metropolitan areas, and especially in the Northeast and in California, ISDN services have become widely available for small business and residential users. ISDN service is now being offered for as little as $30 per month in some places. There is, however, a rather expensive up-front charge for the ISDN modem as described below.

There are two types of ISDN: dedicated and dial-up. Both require some form of hardware that is capable of translating the analog phone signal into a digital signal. One of the better known ISDN modems is the Ascend Pipeline 75, which costs about $1,200. Although expensive up-front, using ISDN offers other benefits besides improved speed. For instance, with a dedicated ISDN connection, you can have a unique Internet Protocol (IP) address on the Internet (as opposed to being under a proxy).

Besides the expense, there is another drawback to using ISDN. It takes some engineering knowledge to get everything hooked up correctly. The cost and engineering aspects of ISDN are what make it a less than desirable candidate for the home user connection to the Internet. However, it is often an ideal solution for the small business or organization that wishes to have higher speed without the investment in 'fractional' T-1 connections (128, 256, 381, and 512 KBps connections), which are much more expensive. So, back to what is practical for home use.

The New V.90 56 KBps Modems

The modem was a major breakthrough in telecommunications. It allowed computers to communicate by converting their digital information into an analog signal to travel through the public telephone network. However, there is a limit to the amount of information that an analog telephone line can hold. At the present time this limit is about 56 KBps. Commonly available modems now have a maximum speed of 56 KBps, but are limited by other elements of the phone system and routinely go no faster than 45 KBps.

Some phone lines do not even support 56 KBps connections at all, and until very recently (February, 1998) there have been two competing 56 KBps standards (the X2 standard from U.S. Robotics and the K56flex from Rockwell/Lucent). This standards problem has now been resolved by the International Telecommunications Union. The ITU, a United Nations agency, coordinates global communications standards.

Their determined standard, designated V.90, is expected to be widely used for applications such as Internet and on-line service access. V.90 modems are designed for connections which are completely digital at one end and feature a digital to analog conversion at the other. Download speeds of 56 KBps and upload speeds of 33.6 KBps are possible, depending upon telephone line conditions. Manufacturers currently producing modems based on either of the previous proprietary schemes have already stated they will rapidly migrate to this new standard, and may offer software to upgrade their previous offerings to the V.90 standard.

Now, while there is a big difference between traveling the Internet on a 56 KBps modem and a 28.8 KBps modem, and while it is an inexpensive and practical way to get higher information transfer speeds to the home, the V.90 approach is not the optimal answer for "The Last Mile Problem." A modem transferring information at 56 KBps-which, as mentioned above, is seldom actually seen due to the imperfections of phone line technology- will still limit realization of frustration-free, multimedia computing over the web. So what other approaches are on the horizon?

T-1's and FTTH/FTTN...Non-Solutions?

The standard in very fast connections for larger organizations continues to be the T-1, with a top transfer speed of 1.5 MegaBytes per second (MBps). The cost of the T-1 line has come down to $1,000 per month in Virginia, and the interconnectors to a Local Area Network (LAN) can be had for as little as $5,000. Many businesses find this is an affordable price, but it is obviously out of the question for the average household.

There is also a school of thought known as FTTH (Fiber-to-the-Home) or FTTN (Fiber-to-the Neighborhood) that argues we should replace all of the copper with fiber-optic cable, which is capable of delivering voice and data at blistering speeds. If we were able to get fiber to the larger communities, we could then use other technologies to deliver the high-speed data into the home. The problem is that the 130 million plus phone lines in the U.S. use 700 million miles of copper pairs. According to a study done in 1987, the cost of replacing all of the existing copper wire would be somewhere in the vicinity of $250 billion. Not feasible.

Cable Internet...The Most Likely Solution

Perhaps the most promising-and most likely-source of providing the average household with a connection speedy enough for frustration-free, multimedia Internet computing is the local cable company. Using cable modems, download speeds of 1.4 MBps and upload speeds of 780 KBps are possible. This is 50 times or so faster than the current dial-up technology in use. The cable companies are starting to implement this service, which requires the household to rent a modem. The cost is envisioned at $30 or so per month (once volume is achieved) to the user with no other hardware costs. The bundling of cable television, Internet access, other entertainment options, and even telephone service is likely as the cable companies realize their market potential.

An ancillary version of such a cable company offering is the High Speed Digital Wireless Cable Modem, in which the link between the cable company-or possibly a telecommunication company or even a Internet Service Provider (ISP)-and the user's location is bridged by radio rather than cable. The user would need an antenna, and perhaps some additional hardware, but the effect and speeds would be about the same as regular cable. Because of the need for the antenna and potential other hardware, the costs would be somewhat higher than simple connection to a cable company's line.

In general, cable company service stands as the most likely approach for satisfactory Internet connection for the average household in the early part of the new century. The wiring is already in place in many locations, and it offers the speed to solve most of the information transfer needs for average Internet computing. However, it won't allow portable access to the Internet the way the current use of the telephone system does, and this could be a very big problem for many. Also, it leaves open the question of Internet service for those parts of the country not served by cable, physical or wireless.

ADSL...The Real Future for Fast Connectivity?

There is one other untried technology that also offers the prospect of a practical solution to "The Last Mile Problem". It is the Asymmetric Digital Subscriber Line (ADSL).

ADSL converts an ordinary phone line into a high-speed digital pipe for ultra-fast access to the Internet and corporate networks while also enabling real-time multimedia services. With download speeds as high as 9 MBps and upload speeds of 1 MBps, ADSL is over 300 times faster than the 28.8 KBps dial-up modems and 70 times faster than paired 128 KBps ISDN.

ADSL exploits the unused spectrum capacity in ordinary phone lines, employing advanced modulation techniques to transmit at the speeds mentioned. In spite of this huge increase in capacity, regular phone service remains undisturbed in theory. That means one ADSL line can offer simultaneous channels for data, voice, and video services. Because ADSL connections are point-to-point, circuits are always up, just as with dedicated lines.

This continuous connectivity frees users from having to dial up every time they connect to the network. And because ADSL isn't circuit-switched, it's unaffected by congestion at the provider's Point of Presence (POP), something that can seriously degrade ISDN performance given the amount of dial-up Internet traffic now present.

All in all, ADSL has the prospect of a terrific solution for residential Internet access. However, it involves using the existing telephone lines in ways that haven't been tried before, and there are some technical difficulties that have yet to be resolved. Testing is currently underway by Bell Atlantic and GTE on the feasibility of the approach. It would require the user to purchase/rent a fairly expensive modem or card type device, but this cost could become very affordable if the approach is adopted in volume.

Other Slowdowns

The Internet knowledgeable reader might observe, at this point, that I have not discussed factors beyond (or more accurately, before) "The Last Mile Problem" that affect the speed of Internet interactivity. The explosive growth of the Internet has led to some bottlenecks that impede the overall system's speed at various times of the day. One of the best known of these is the infamous MCI facility in Washington, D. C., where European Internet traffic and most of the continental North American traffic is merged. There are also numerous instances of malfunctioning routers and other hardware glitches on the Internet backbone. In concert with these problems is the reality that many servers are inadequate to the level of activity required of them, leading to slow responses to inquiries.

All in all, there are many elements that slow Internet computing prior to the arrival of the signal at "The Last Mile." However, the vast support provided the Internet's growth by very large and well-heeled technology companies is quickly overcoming these problems, and it appears that taken as a whole they will not constitute a chronic impediment to signal speed in the future.

Conclusion

On the residential end, over the next five years the various competing approaches to increased speed for household Internet connectivity will sort themselves out by price and quality. The consumer can expect one of these approaches-or one totally unseen at this point-to attain supremacy and offer download and upload times that will render interaction speed a non-issue. The price for such service will probably settle in the $25 to $50 per month range, at current dollars. With such service in place, distance learning and commerce in general on the Internet will become second nature for much of the population.

Dr. Leventhal's company, Blueridge Internet Technologies and Services, Inc. (BITS) in Charlottesville, creates and hosts sophisticated web-sites for over 70 organizations in the Central Virginia region. BITS web-site address is: http://www.bitlink.com


Back to Background Materials