Liebhold.html

CFP'93 - Liebhold

CFP'93 - Alternative Telecommunications Architectures for the National Information Infrastructure

by Mike Liebhold

Media Architecture Research
Apple Computer, Inc.
3/1/93

A National Information Infrastructure extends beyond a fiber backbone, through a variety of wired and wireless/broadcast technologies to enable greater accessibility and to provide both fixed and mobile communications and computing.

Based on proposals by the EFF (Electronic Frontier Foundation, Mitch Kapor Exec. Dir.), Rep. Edward Markey is planning the introduction of legislation [1] calling for an "Open platform service ... (to be) ... widely accessible throughout each state ... available at affordable charges based on the reasonably identifiable costs ... available to all customers on a single line basis upon reasonable request." The legislation would further require the FCC to establish regulations within one year, and local carriers to file tariffs within 120 days thereafter.

Impact

The legislation would compel a standardized narrow-band (2B+D) ISDN [2] to be offered throughout the US wherever existing switching equipment would support it (approximately 54%). It is not clear that Congress could actually require the remaining 46% of the telephone exchanges to purchase ISDN equipment or any equipment at all. One interpretation of the language could mean that other equivalent or superior services (e.g. ATM) could be offered instead.

The impact on the NII is not clear. If the RBOCs actually deploy the service and if customers actually buy ISDN compatible equipment, then as critics claim, any funds spent on ISDN would, indeed, delay the ultimate implementation of a broadband (ATM) NII. What the RBOCs would actually do if this bill passed is anybody's guess.[3] Aside from EFF and the switch vendors,[4] there is no real demonstrated demand for ISDN services. Apple introduced an ISDN card, three years ago and then dropped it because of lack of customer interest.

Arguments

Alternative Technologies There may be reasons to consider the Markey/EFF proposal if it were amended to provide equivalent support for the earliest market proof and deployment of other new, and potentially more marketable, higher performance network technologies including:

1. High bandwidth wireless communications services

There are many new wireless technologies promising useful access to the NII:

Apple Data PCS - User provided carrier services offer tremendous potential for college and universities.
Cellularvision - a new two way microwave television system developed by 'Suite 12' and demonstrated in Brooklyn N.Y. operating at 28 Ghz.
ADML - Asymmetrical Digital Microcell Line - Bellcore is reported to be investigating delivery of 50mbit/sec channels to consumers at frequencies up to 70 Ghz.

2. Hybrid cable/telco networks

The legislation should also be amended to provide genuine incentives for local regional telcos and cable operators to cooperate in building hybrid digital networks [5] capable of providing many more promising NII applications of fully interactive multimedia - (remote learning, distributed medical services, video lectures on demand, online simulations). These networks will combine the best properties of both telephone and cable network topologies: Telephone networks offer reliable two-way interactive, addressable functionalities, but lack sufficient bandwidth to support digital multimedia. Cable television systems, on the other hand, offer tremendous broadband multimedia capabilities, but lack both the reliability and two-way addressability of the telephone system.

3. High Performance Cable Networks

New 'Fiber to the Node' cable television networks offer outstanding potential to become fully interactive broadband delivery media for the NII. Currently, cable television capital improvement spending on fiber optic cable is significantly greater than any telco's investment in lower performance ISDN. According to industry newsletter,[6] 22 cities already have 'Fiber to the Node'[7] systems in place. According to Cable Labs, 45 regions may have 'Metropolitan Regional Fiber Hubs' - interconnecting various 'Fiber to the Node' head-end facilities using multiple gigabit ATM/SONET fiber rings for shared access switched access to long distance services, multimedia servers, and satellite terminals.

Time/Warner Corporation has demonstrated in Queens NY. a 'Fiber to the Node' system which delivers communications resources capable of 500 television channels along with basic telephone services. Several CATV companies in Pennsylvania have agreed to form a statewide fiberoptic network linking together a number of 'Regional Hubs' capable of provide a wide range on services including education, and healthcare along with entertainment services.

4. ADSL & Fiber-in-Loop

The RBOCS ( Regional Bell Operating Companies) - are conducting many private trials exploring a wide range of services on a variety of network topologies including video over ADSL [8] at 1-6 megabits/sec, FITL (Fiber in the Loop) ATM is particularly promising for private high performance, multimedia networks. All of these offer considerable more promise than ISDN for the delivery of the kind of interactive multimedia services that many believe to be prerequisite for financial success of a new communications media. (see following)

Demonstrated Value

2B+D= 2 64 kilobit channels and one 16 kilobit channel over a copper pair has not been demonstrated as adequate for interactive multimedia.Many believe consumer acceptance of new network services will depend on the higher quality experience that multimedia delivers. Few, if any compelling applications have been demonstrated for ISDN. Neither customer demand nor market value have been credibly demonstrated.

Cost vs. Broadband

The numbers are inconclusive. Tim Regan of Corning Inc. cites Consumer Federation of America estimates cost of an ISDN upgrade to be $30-60 billion.

Some ISDN proponents argue that a complete fiber optic infrastructure would cost $450 billion. Corning Inc. disputes those claims. They project incremental cost of deployment of an interactive broadband fiber net by 2015 to $23 billion. Costs would be born almost completely by private industry. Fiber has almost achieved cost parity with copper in on cost/performance basis. So, there would be no or little incremental costs. A fiber infrastructure would free on that basis if we're prepared to wait until 2030-2040 for universal deployment. If we want a broadband infrastructure by 2015 there is an incremental investment of $23B (which compares well with the costs above) for a copper ISDN upgrade.

Conclusion

1. The Government investment in network services should support technologic diversity while providing incentives for interoperability.

Mike Roberts, Exec. Director of Networking for Educom argues that the NII must remain technically neutral. Educom is the association of academic computing and communications. Mike says" Let technology experiments flourish and let demand determine the appropriate technology..." There are just too many other promising competing technologies. Including digital Cable TV TCP/IP and ATM over fiber, and others. Given the same early support proposed for ISDN, these networks offer much greater potential for the NII access.

2. An investment in ISDN will actually delay the early deployment of a high performance NII.

Ken Phillips, Citibank Fellow says:

"ISDN could easily become the major impediment to the US NII because, if the local exchange companies proceed with the installation of central office equipment, into today's depreciation environment, then Telco will be unable to make timely replacements of this technology with more modern media capable of supporting new compelling applications."

The key technical issue for the national information infrastructure is how to manage concurrent change in computers, telecommunications, video, and publishing technologies which enable a near-infinite array of integrated networks and multimedia services to be developed.

References

1 Draft proposal is called "The Telecommunications Competition and services Act of 1992". (back to text)

2 ISDN = Integrated Services digital Network. 2B+D = 2-64 kilobit "Bearer" user channels and 1-16 kilobit "Data" signaling channel over a copper pair- also called BRI (Basic Rate Interface). (back to text)

3 After ten years, ISDN is technically already "on the price list" in a few exchanges. In actuality it is nearly impossible to obtain a connection. New England Bell, for example offered the service only reluctantly after considerable pressure from EFF. According to Richard Solomon, MIT, New England Bell still makes it extraordinarily difficult for an ordinary person to actually obtain an ISDN hook-up . It could lead one to believe that the telcos don't really believe in ISDN. (back to text)

4 One RBOC told me that AT&T has installed ISDN switches on spec. That is, they agreed to collect a percentage of transactions instead of direct payments from the local carrier. (back to text)

5 Hybrid Cable/Telco networks are precisely what the _First Cities_ project intends to develop. First Cities is a joint effort of 14 or so companies chartered to: 1) to develop technologies, 2) prove the applications, and 3) to prove the business case for _interactive multimedia networks._ The basic assumption is: A network capable of profitable delivery of commercial multimedia services will also be an ideally suited and affordable medium for public services for schools, hospitals, libraries, and museums. First Cities member companies include MCC, Apple, Bellcore, Eastman Kodak, Kaleida Labs, Tandem, S.W. Bell, U.S. West, Corning, and others... (back to text)

6 _Cable-Telco Report_, 6/92 (back to text)

7 Where fiber has been extended to serve 2000, or fewer households. (back to text)

8 ADSL - Asymmetric Digital Subscriber Loop (back to text)