FreeWire

A New Year And Some New Challenges

by Bill Frezza

As you have probably read, Network Computing is expanding to 20 issues this year--a testament to the growth of the industry. This provides a wonderful opportunity to delve more deeply into favorite subjects as well as explore new ones. I intend to do both, continuing my coverage of the struggling wireless data industry while branching out into what I am convinced will be the decade's most significant trend in computing and telecommunications--Public Network Computing.

Fellow columnist Timothy Haight, in the August 1, 1995, issue, identified Public Network Computing as the fifth wave in the evolution of the computer industry. The first three waves have become the stuff of history--mainframes, minicomputers and PCs. Mark Stahlman, former analyst at Alex Brown & Sons and former columnist for this magazine, is generally credited with identifying the fourth wave--network computing--forecasting the dynamic growth of workstations as they broke from the PC model and the LAN as it became the essential nervous system of corporate America.

Public Network Computing--the extension of distributed computing outside the enterprise conjoined with networks of other public and private entities--is ready to take center stage in the lives of the people who read this magazine. Public Network Computing will be as different from today's telecommunications and computing environments as PCs and servers were to dumb terminals and minicomputers.

The first thing you'll notice is that the rigid demarcation between pu blic and private facilities will blur as IS managers reach across town and across continents with new network configuration and control capabilities. As this happens, the transport elements of the public infrastructure will be forced to unbundle, after which they will fragment, commoditize and then take off on a new round of exponential growth. Simultaneously, a new driving factor and large element of uncertainty will be introduced into the mix, namely the consumer. Long fenced off behind antiquated local telephone networks, this PC- and credit card-equipped change agent will charge onto the Public Network Computing scene just as corporations are trying to find new ways to reach both customers and their employees.

Creative Destruction Squatting in the middle of all this is the venerable Public Switched Telephone Network (PSTN). The PSTN is a highly tuned, monolithic, application-specific switching system that today is tightly bundled with enormous quantities of generic transport. Its design principles evolved over the course of 100 years, the last 60 of which occurred in a rigid and protected market. Most important, the massive central office switches that control call processing, routing and billing--special purpose mainframes, really--are architected around underlying traffic models that have barely changed in decades.

Similarly, the business models that support the financing, deployment and operation of the PSTN have been frozen by Public Utility Commissions. Their main purpose is to protect and preserve the status quo while pandering to their political constituencies, namely the local Bell Operating Companies and their stockholders. Investment decisions are based on large upfront capital expenditures and guaranteed long-term cost recovery. This is accomplished in highly mature markets with no competitive uncertainty, no substitution of services and no obsolescence that isn't carefully planned. This economic model has had an enormous impact on both system architectures and corporate cul tures.

Truth be told, the PSTN is very, very good at what it was designed to do, just as mainframes did an admirable job of batch processing the payroll. The fact that the phone network can do anything else at all, like support fax and circuit switched data, is a complete accident made possible because this other traffic was disguised to mimic the parameters of a voice call. This is hideously inefficient, but if the computer industry had waited for the phone companies to figure out how to provide data services rather than go ahead and invent analog modems that are really voice-call spoofers, we would all still be computing in splendid isolation.

There is actually an important lesson here. The more the phone companies try to hold back the tide of Public Network Computing, forcing customers to live with bundled switching and transport, as well as tariff structures and business models that are hopelessly out of date, the more surely phone companies will be rendered obsolete. Just watch the major Internet Service Providers (ISPs) like PSI, UUNet and Netcom frantically laying down track, building depots and switching centers in every city and town. The traffic that their local Points of Presence (POPs) are bleeding away from the PSTN may not represent much volume today, but it could account for most of tomorrow's growth.

Oddly enough, the underlying transport the ISPs use is exactly the same transport used by the phone companies--after all, a DS3 line is a DS3 line. The difference is that this transport is configured to provide application-independent services with a high degree of modularity and a presumption of distributed intelligence. None of the applications that the ISPs are chasing are mature, forecastable or readily characterizable. (Heck, many of them probably aren't even real.) But, flexibility allows them to bet on the aggregate. Both the ISPs and Competitive Access Providers (CAPs), to a certain extent, are learning to optimize business against a moving target with no guaranteed rate of return, no decades-long cost recovery and every possibility that a substitute service can come along tomorrow and take their business away. Just think how fierce this makes both their system architectures and their management team.

Now, imagine these two together in a competitive market. In the Darwinian struggle that looms ahead, the PSTN can be compared with the marsupials of Australia: highly evolved to fill a protected environmental niche. The Internet is a wild predatory mammal, let loose by a massive wave of immigrants that have little knowledge or concern for the native flora and fauna. Take a wild guess which is going to kick butt when the last barriers to competition come down.

Ah, but can't the PSTN always retreat to its lair and live off voice traffic? Don't be so sure about that. Internet phone is being scoffed at as an oddity by the phone companies, but you can bet that at the very least it will massively destabilize the international long distance tariff structures. Only an idiot is going to pay $35 an hour for international long distance a few years from now. Network arbitrage--the ability to u se application-independent switching systems on top of commodity transport to undercut application-specific bundled and tariffed services--will become a big game in the age of Public Network Computing.

Summing It Up So, putting it all together, what are the top 10 defining technical and business characteristics of Public Network Computing? 1) Competitive yet compatible application-independent switching connected with commodity transport; 2) Distributed intelligence flexibly accommodating ever-changing traffic models; 3) Modular deployment aimed at hot markets, unencumbered by social mandates to guarantee "Universal Service"; 4) Incremental provisioning with lower first-costs, shorter amortization periods and a rate of capital equipment turnover closer to the PC industry than the telecom industry; 5) Destabilizing network arbitrage that will do more to dissolve the Public Utility Commissions' cozy little club than any act of Congress; 6) A high rate of new market entry, and a correspon dingly high rate of mergers and failures; 7) Massive trauma and angst as the local Bell Operating Companies are forced to cut the dividend, break the bloated unions and toss out their old-boy, brain-dead management as they struggle to reinvent themselves; 8) A mind-boggling array of choices for businesses and consumers; 9) A massive increase in aggregate telecommunications traffic; and 10) Virtual integration of selected pieces of all of this under the control of the enterprise. What a great time to be a Networkologist!

So boot up, tune in and let's watch the game. (I forgot to mention that the advent of Public Network Computing will also provide full employment for cheeky commentators.) While I will continue to bring you FreeWire in each issue of Network Computing, I have also begun writing biweekly op-ed pieces on technology, culture and politics for our sister publication, CommunicationsWeek. In addition, CMP Interactive recently launched a Web service called TechWeb Gurus in which I am participating. You can visit us at http://techweb. cmp.com/gurus/gurus.html and put in your two cents. I've posted a wireless data directory online so you can get phone numbers and Internet contact information for any company ever mentioned here. I will also be leading an ongoing series of interactive debates and dialogue on issues central to the industry.

So come on by. As the saying goes, I'll see you on the Net.

Bill Frezza is the president of Wireless Computing Associates. He can be reached via e-mail at frezza@radiomail.net or on the Web at techweb.cmp.com/gurus/ gurus.html.

January 15, 1996