Friday, December 03, 2004

The Pennsylvania 180°

Fast Internet Service for The People (

In an interview in the current (December/January) edition of queue magazine, Alan Kay, one of the most influential figures in computer programming over the last 30 years or so, commented on the pressures that commercial pressures place on technology developers. Kay's specific comments focused on the environment in a places like the Xerox Palo Alto Research Center (XPARC) versus the pressures to produce products in commercial entities. In the interview (which is not yet available online), Kay is talking about how he believes the marketing people at Sun Microsystems rushed the release of the programming language Java and, as a result, key work of the software engineers was frozen in place. The result is a programming language that, in Kay's view, does not live up to its promise of running the same on any and all computing platforms.

What Kay is talking about is the pressure of deadlines in market environments (even politics) which is completely different from deadlines in a research environment. The effect of that pressure can be hasty decisions which carry with them long-term implications.

This story is relevant to understanding what has transpired in Pennsylvania in the past few days. There, under pressure of a powerful telecom lobby and an impending signature of a law by the governor there, the City of Philadelphia, in effect, sold out the rest of the Commonwealth in order to get to deploy the broadband wireless network which prompted this telecom onslaught.

The core story is that RBOC Verizon attempted to use its influence in that state's legislature to prevent Philadelphia from deploying a city-wide broadband wireless network. It is, apparently, a page from the standard playbook of how, early in the 21st century, telecom giants are trying to preserve their monopolies on infrastructure. This is exactly the tact that BellSouth and Cox took in Louisiana to try to prevent Lafayette Utilities System (LUS) from deploying a fiber to the premises plant here. It is also the approach being used by Verizon and other RBOCs in other states to try to squelch the growing number of local governments that have decided that their historic role as providers of essential infrastructure should now include the infrastructure of the information economy: networks.

On one level, Verizon failed, because in last minute bargaining with Governor Ed Rendell, the City of Philadelphia won the right to deploy its wireless system — if it can do so by January 1, 2006. Any other municipal system not in place or in the process of being deployed by then will have to get permission from the local incumbent carrier to deploy such infrastructure, according to this story on the final legislation.

While the new law is not a complete victory for Verizon, even the New York Times notes that the law will make it more difficult for other municipalities in the Keystone State to deploy new network infrastructure there. What that does is leave the decision as to when these communities to gain access to new network infrastructure in the hands of Verizon and, I guess, cable companies like Comcast.

This is not the status quo, because now there is a body of law that will enable the incumbents to stop municipal network efforts without much effort. It represents a new tightening of the grip of telecom giants on essential infrastructure in Pennsylvania. It also marks a dramatic reversal of fortunes for Verizon and Pennsylvania from earlier in this decade when state regulators nearly separated Verizon's antecedant Bell Atlantic from its network.

To understand how breathtaking a defeat this is, one has to know a little history about telecom regulation in the Commonwealth of Pennsylvania, dating back to the heady years of the late 1990s when it seemed just about anything was possible.

In those days, Verizon didn't exist. It's a company formed by the merger of NYNEX, Bell Atlantic predecessor, Bell Atlantic, and maybe another company or two. Bell Atlantic was the RBOC providing service in Pennsylvania dating back to the breakup of AT&T back in 1984.

With the passage of the Telecommunications Act of 1996, the RBOCs were obligated to open their network infrastructure to competitors. The idea of welcoming competition — and over their own network at that! — ran counter to the way the phone companies were hard wired. The RBOCs were the products of a national monopoly which had been granted regional monopolies when the national monopoly was broken up. Monopoly behavior is in their corporate DNA. Getting RBOCs to open their networks and work nicely with competitors was more than teaching an old dog new tricks; it was like trying to teach a dog to meow!

Regulators at the Pennsylvania Public Utilities Commission recognized something more fundamental, though. They recognized that the heart of the problem was that the RBOCs were being asked to be both network operators and service providers. The network operator (infrastructure) side of the company had to play nice with competitors, welcoming to the network and helping them connect to customers. The service provider side was waging a war against those same competitors, fighting to maintain every percent of that near 100 percent market share the RBOCs had.

The Pennsylvania PUC recognized that the new line sharing requirements forced a form of corporate schizophrenia on the RBOCs resulting from the vertical integration of those companies. That is, the RBOCs were (and still are) organized along the principle of owning the infrastructure and delivering the services to customers. The line sharing requirements dictated that at least one part of these vertically-integrated companies act as if the industry was horizontally segmented; that is, that infrastructure was one element of the industry, that service provision was another.

At some point around the turn of the century, when it became clear that the RBOCs could not really resolve this inner conflict and play nice with the new competitive local exchange carriers (CLECs), someone at the Pennsylvania PUC proposed a radical solution: force the RBOC to divest itself of ownership of the network.

The Penn PUC came damned close to authorizing a full divestiture of Bell Atlantic's network infrastructure in this decision (it's a Word file) on interconnection agreements. Instead, the PUC chose "structural separation" which had the effect of walling off the network side of Bell Atlantic from the service selling side. Apparently, the near-death experience of having its network taken away from it was enough of a scare to reform Bell Atlantic's behavior towards CLECs.

The Penn PUC's decision to even consider network divestiture was a decision closely watched by regulators and competitors across the country. Had the PUC decided to order a divestiture, it surely would have resulted in a lengthy legal fight. Nonetheless, it would likely have fundamentally changed the terms of the discussion of telecom today.

Viewed in the light of this history, the bill that will legally cede network infrastructure control to Verizon and cable companies constitutes a tremendous fall for businesses and consumers there. It will be interesting to see how this plays out over the next couple of years. Clearly, Verizon has demonstrated that its political influence in Pennsylvania has risen considerably since earlier this decade when the PUC threatened to take it's network away.

What a turn around!


John said...

Mike, thanks for a great post. First, Kay is one of my all-time heroes. But the great part far me was the background on the Pennsylvannia Public Utilities Commission. That was priceless. And, of course, deepens my depressive reaction to the Pennsylvannia story itself. To know that Pennsylvania came so close to rational behavior...

There has also been some disturbing discussion on the net about other aspects of the Pennsylvannia telecom bill. The part of it that gives transnational corporations control over local communities' infrastructure decisions is the one we'd be most in tune with here and seems to be getting the majority of the national reaction. But there is more, including a big "incentive" (read: "tax giveaway") to encourage Verizion to build out broadband; a deal that Verizon has already welched on once before in Pennsylvania. Some commentators say that the earlier cheat cost each Pennsylvannia household 782 dollars. And we thought that the days of the robber barons controlling local legislatures in the US were over. Apparently not.

Mike Stagg said...

Yes, John, the subsidy for network expansion by these corporate giants is adding insult to injury. After all, what the Pennsylvania Legislature is saying to local communities is "We and our patrons at Verizon and Comcast know what's good for your community better than you do."

As for the PUC aspect of the story, I was a member of the Rural Policy Research Institute's Telecommunications Research Panel in the late 90s and early in this decade. We were looking at telecom issues all over the country. We had intense 'discussions' about the role of infrastructure and who should build it and own it.

The most impressive presentation I ever heard on this topic, though, came from William Crowe, the CEO of Level 3 Communications at the SuperComm conference in Atlanta. I think it was in 2001.

He went into great detail about this notion of telecom being a vertically integrated industry and how that orientation had stymied and would continue to stymie innovation. He pointed to the computer industry as proof of his thesis. He noted that when IBM dominated the industry as a vertically integrated provider of hardware, software and services, the industry was stagnant. Crowe claimed the computer industry exploded with innovation only after it began to orient itself along a horizontal axis.

The examples he gave were that companies began to dominate segments and, in the push, to maintain their market leadership, used innovation to deliver value to customers. OK, so it breaks down pretty badly when you look at Microsoft's domination of operating systems (;-0), but otherwise it makes sense.

What the Pennsylvania PUC was trying to do was make network sharing work as a practical matter. The inability or unwillingess of the RBOCs to allow that sharing to take place drove the PUC to view the vertical integration of Bell Atlantic (and all RBOCs) as the core of the problem.

Kay, in his interview in queue, makes the same point about OSs. He notes that innovation in programming and in interfaces has slowed dramatically since the market has come to be dominated by Microsoft. He doesn't say that with any animus towards Microsoft, but as an observer of the industry over a four decade period.

Dominance invests the owner with a huge stake in the preservation of the status quo. It's all about milking the existing situation for all its worth by adding features without improving the core of the system. Kay calls this approach the "agglutinative approach" to programming.

The vertically integrated phone system, owned first by one company, then by about nine (except in rural areas), now down to about four, did very little innovation for decades. There was a flurry of innovation driven by competitive pressures in the years since passage of the Telecommunications Act of 1996, but that will likely lessen as the telecom companies tighten their grip on infrastructure.

The one open door to innovators — even in an era of incumbent dominated infrastructure — will be the Internet carried by fat pipes (eventually) to every home and business. This is tenuous, because those companies can close ports in their network which can have the effect of blocking certain types of services. Some phone and cable companies have done this with VoIP service providers like Vonage.

It is this open Internet fat pipe in the LUS plan (remember LUS?!) that makes me a supporter of their proposal, despite the fact that they don't want to open the fiber itself to other voice and cable providers. Ultimately, those won't matter. The phenomena of the Internet Protocol as a wrecker of one-to-many business models will take care of that.

My prediction is that within five to seven years of the completion of the LUS fiber to the home buildout, more revenue will be earned through the sale of bandwidth than through cable or voice services.

So, as long as there is an Internet connection over a fat pipe, there is the possibility for innovation and that is what will enable the horizontal orientation of the industry to come into being.