VIII. Future Paths to Connectivity
We will separate the discussion of this section into two parts. First we will talk about technologies for the local loop, since we have already identified this element as the key to affordable school networking solutions. Then we will talk about the services which can be provided over this local loop, indicating the range of bandwidths available through each service and commenting where possible on the likely pricing of these services.
For sites with low-speed SLIP connectivity the simplest improvement in bandwidth will come from advances in modem technology. Current modems provide raw data rates of 14,400 bps (enhanced somewhat by data compression technology to give effective data rates two or three times higher). The next generation of modems, which is just now coming to market, will double these data rates.
The signaling technology used for leased lines and Frame Relay is known as DDS. Since this layer of technology is hidden from view in the line adapters (DSUs or CSU/DSUs) used to connect to the leased line, we need not discuss this technology in any further detail.
For unconditioned copper we have mentioned the use of Short-Haul Modems to implement 56,000 bps links. A newer technology, known as HDSL, allows one to provide much higher bandwidths over unconditioned copper. This allows for T1 speeds (1,544,000 bps) at costs comparable to those of analog voice lines. The limitations of this service have to do with the distance of your site from the phone company's central office and the availability of this type of service in your locale.
Underlying ISDN service is a signaling technology known as 2B1Q. The advantages of this technology are that it exists and is in the current deployment schedules of most telephone companies in the country, although these schedules often extend uncomfortably far into the future, especially where inner city or rural schools are involved. It is also inexpensive, and aside from the problems caused by tariffs which add connection charges for the service, it is a good current option for many sites.
As mentioned previously, there is a brief aside that we want to offer with regard to tariffs which add a connect-time charge for ISDN data connections. Sometimes these tariffs allow ISDN voice calls to go through without a connect-time charge. It turns out that the difference between ISDN data connections and ISDN voice connections is very minor. Both involve digitized signals. The voice connections use only 112,000 bps of the available 128,000 channel. But one can actually pass data over a so-called ISDN voice connection. You can buy the hardware which allows you to do this from independent suppliers or from the telephone company itself. For some school districts this may be a very economical approach. In the long run one suspects that the telephone tariff structure may be rationalized and the connect-time surcharge for ISDN data calls simply eliminated.
So far we have been discussing local loops which can be provided over the standard twisted pairs of copper wire traditionally used for telephone hookups. It is natural to emphasize this option, since there is so much of this wire already in place. But new applications may require new wiring schemes, not just for our schools but for our whole community. Let's look briefly at what's on the horizon here.
A great deal of fiber optic cable is presently being installed. Most of it is being deployed as part of the network backbone, be it the telephone network, the cable TV network or the Internet itself. The possibility exists of running fiber optic cable direct to the schools. Several technologies exist for exploiting this development. A service known as FDDI can provide fixed speeds of 100 million bps. Another, scalable, technology known as SONET carries the available bandwidth even higher, with speeds ranging from 50 million bps up to several billion bps.
Before one considers fiber optic links, which require line adapters and computer interfaces which are currently quite expensive, it's worth asking what other physical means of connectivity might already be in place. An obvious candidate, if we are looking simply at available wires, is the coaxial cable that carries the television signals for many communities. Technology exists to turn one cable television channel or a channel pair into a data channel. A single channel can provide data rates of 4.5 million bps, and a pair as much as 10 million bps, which is as fast as the school LAN that we have recommended at the start of this section. This is potentially a very inexpensive and attractive solution for school connectivity. The principal drawback has to do with the reliability of the cable system. The standard of performance for data transmission is far more stringent than the standards for television transmission. Hence it remains to be seen if the cable option can provide a reliable enough mechanism for a school network which could become the basis for all routine school services. As cable companies convert portions of their physical plant to fiber, and as they gear up for their enhanced role in the National Information Infrastructure, they may be able to deliver network services economically and reliably through this mechanism.
If we are inclined to find fault with some of the wires available for data transmission, how about giving up on wires entirely? Wireless transmission technology is another area undergoing rapid development. Examples include transmission via laser beams, microwaves of radio waves. Lasers and microwaves require an unobstructed line of sight between the points to be connected, so their application is somewhat specialized. A problem with radio transmissions is that the radio spectrum is limited, so it's natural to place a premium on its use. Cellular telephone service is thus far more expensive than service from fixed, hard-wired phones. Special frequency allotments for school use could solve this problem, as of course could special tariffs for fixed-line services. This is a technology to watch, but one shouldn't put off local efforts at school networking until wireless technology matures; it may turn out not to be the best solution for school connectivity.
For any of these physical links there is a range of services that one can contemplate. We are interested in packet-based services which can operate over the high-speed links that we have been discussing. Such services are known collectively in the telephone business as fast packet services. We list below the services which appear most promising at the present time and most likely to show up on the menus of your local network access providers or telephone companies.
Frame Relay. We discussed this in some detail in section VI. It is a fast-packet service, available in speeds ranging from 56,000 bps to 1,544,000 bps. Pricing is competitive with that of leased lines, but Frame Relay wins as a scalable option. If distances to school sites are large, Frame Relay will be cheaper than leased lines, since its provision is typically independent of the distance from the telephone company's central office.
SMDS. This is another fast-packet service. Speeds range from 1,544,000 up to nearly 45 million bps.
ATM. This is a technology which is being developed very aggressively on the national scale. It is based upon the transmission of fixed-size cells through a process known as cell relay, a mechanism which is particularly suitable for the integrated transmission of voice, video and data. ATM could easily provide the underpinnings for all of the future Internet, as well as a high-speed option for local connectivity. Available speeds range from 30 million bps to several billion bps.
The survey of the present section is necessarily incomplete, since there exist many new technologies currently under study, and it is hard to guess the rate of deployment even for those technologies which have been publicly announced. We hope that our remarks provide a useful guide to the jargon of the field if not to the content on the deepest level. Further details of the jargon are contained in the glossary which follows in section IX.
We conclude our survey with a mention of technologies which don't quite fit into our general framework but which may become widely deployed. We have emphasized strongly the value of a symmetrical connection for school sites on the Internet. For home use, in particular, there may be more demand for asymmetrical connections, in which more data flows to the home than from the home. An obvious application is that of video on demand, a topic of current intense development efforts. Bell Atlantic has recently announced an asymmetrical service called ADSL (where the "A" stands for asymmetric). This service can provide very high bandwidth in one direction (to the customer's house) with a much slower back channel. This sort of technology was designed for applications such a home video on demand. It could be quite nice for browsing remote libraries with text, graphics, sound and video clips. But it could not be used for a school site to disseminate such a library, since the back channel would probably be too slow for anything but text and simple graphics.
Another asymmetric technology has been proposed for cable TV systems. As described by the Hybrid Corporation, this technology provides 10 million bps in the forward direction, meaning toward the customer. The reverse channel requires a separate telephone link. Technicians would qualify this arrangement as "clunky," since the separate channel doubles the expected failure rate, but this approach does allow for high bandwidth in one direction, albeit over the same cable TV system whose reliability we have called into question earlier in this section.
We have tried to be fairly critical with each of the options discussed in this paper. Practical experience shows that we are probably being more critical than is strictly necessary. The evolution of a National Information Infrastructure will probably provide schools with a broad range of acceptable options for school network connectivity. We hope that the present paper will enable the schools to be intelligent customers in their dealings with network providers, telephone companies and various hardware and software vendors. Some of the hard questions which school people might ask in their dealings with these suppliers could save them significant amounts of money and perhaps ensure that students and teachers in their school will be provided with a network which functions as networks really should - seamlessly and largely invisible.