dgc@LOCUS.UCLA.EDU.UUCP (12/24/86)
The articles about ISDN (Integrated Service Data Network) are appearing fast and furious. Yet I have seen little discussion of its potential impact. It appears that the service will appear, to the small user, as two 64 kilobit-per-second (kbps) full-duplex channels and one 16 kbps full-duplex channel on each ordinary telephone line. One of the 64 kbps channels would normally be used for voice, although it could be used for data. The 16 kbps channel would be used for both signaling (presumably replacing touch tone, etc.) and data while the other 64 kbps channel would be entirely for data. For large users, many more channels would be provided on one pair, but these would be compatible with the above. Though the exact configuration has yet to be determined (part of the standards-setting still taking place -- the delay is apparently due to the differences between the North American and European networks). The telephone companies (BOC's and long distance services in the USA, PTT's in Europe, etc.) would provide the switching for these services. It seems that if this service were provided at a reasonable price and if some reasonable sort of device that would permit RS232 to connect to these channels were provided, then the current "audio-frequency" modems would soon be obsolete. More than that, if the interface chips were widely available and reasonably priced, the standard ISDN interface could easily come to replace RS232 as the standard terminal interface. Advantages would be the higher speed that could be used and a more modern protocol (though I have yet to see the detailed specs). However, judging by past experience, one might expect the telephone companies to "kill the golden goose", most likely by so overpricing the service that modem users won't want to switch. For example, one might expect that there would not be a flat-rate service (given the opposition to such services by BOC's), such as that which is currently provided to many modem users. Even a "small" charge (5 cents/minutes) could mount up quite rapidly. There are many other potential pitfalls. With any of them, the service could die and we could be stuck with the present system for another 25 years. How should such a service be priced? What kind of performance guarantees can we expect in terms of error rate, reliability, and delays? What kind of protocols will it use? Will control be as easy as the present system, which can be handled by simple intelligent modems (Hayes compatible, etc.)? Will there be a wide availability of the interface chips and will these be useful for direct terminal to computer connection (without going through the telephone system)? Will services such as TRW credit (which use automatic dialers and modems to validate credit cards) quickly switch to the new service? Will it replace slow-speed networks such as arpanet (why share the 64kbps arpanet when you can have your own direct line just as fast)? Wil there be "airline" style pricing where, highly-competitive, heavily travelled routes, e.g., Los Angeles-New York have much lower prices than much shorter, lightly travelled routes? (This hasn't happened yet on long-distance services, though I don't know why.) dgc David G. Cantor Internet: dgc@cs.ucla.edu UUCP: ...!{ihnp4, randvax, sdcrdcf, ucbvax}!ucla-cs!dgc