Wednesday, September 22, 2004

Exhibit showcases how fiber optics works

The Daily Advertiser has posted a story on the ADC fiber trailer which stopped in Lafayette on Tuesday. John and I made it there for what we thought would be a brief visit. Almost two hours later, we went our separate ways, carrying with us more questions than we had when we arrived.

We spoke with Pat Sims of ADC who also happens to be a member of the Fiber To The Home Council, a trade group of companies involved in various aspects of this national technology infrastructure movement.

Mr. Sims was a wealth of information and tolerated our many questions focused on various approaches to fiber network architecture and configuration.

One of the major things that I took away from this session is the confirmation of points we've made with varying degrees of effectiveness here: specifically, that EVERYTHING is migrating to Internet Protocol-based services — voice and cable video included (data is already there).

The source for most of the new questions we have is the information contained in the draft feasibility study released by LUS last week when viewed from the perspective of where the technology is heading versus the direction LUS hints that it is heading in the pages of its feasibility study.

The session with Pat Sims confirmed my fears that the LUS team is not looking far enough "down the field" to see where the technology is heading. Or, perhaps worse, sees where the technology is heading but does not recognize the implications of that on network architecture and bandwidth usage.

The danger from this, as I see it, is that LUS might deploy a compromised system that will have more limited value to the community than might be feasible if a clearer-eyed view of trends were applied. It seems to me that LUS has been paying too much attention to BellSouth and others at the expense of heeding Carter Mead's advice of "listening to the technology." Mead has played a critical role in the development of digital technology. You can learn more about him in George Gilder's 1989 book Microcosm: The Quantum Revolution in Economics and Technology, which details the history of the science which led to the development of the microchip and assesses the impact of things like Moore's Law.

Again, speaking only for myself, this concern is compounded by the fact that it appears that the LUS plan is not going to meet the rough deployment timeline which LUS Director Terry Huval laid out for the Consolidated Government Council and the public in a hearing before the council in June. At that meeting, Huval said that it was possible that LUS could beginning to deploy its system as early as the end of 2005. That no longer appears to be likely, as now appears that LUS will not even present its plan to the Council for action until October — at the earliest.

Based on the construction schedule of iProvo in Utah, rollout of the entire system could take at least 18 months. So, it is reasonable to assume that LUS is three — maybe four — years away from full deployment of its fiber to the premises system. The applications that will drive bandwidth demand at that point in the future probably don't exist now; they're being developed or are being tested in other fiber-rich environments like South Korea or Japan.

What is worrisome about the information contained in the draft feasibility study is that LUS appears to be inclined to deploy a system that is better suited for a provider seeking to preserve investments in legacy technologies, such as that of telephone or even less advanced cable companies. I heavily qualify this statement because the draft feasibility is a work in progress, so we don't really know what the details of the LUS proposal will ultimately look like.

For instance, in its feasibility study, LUS says it will buy a telephone switch. With Cox, BellSouth and other providers having announced their intentions to (eventually) migrate all of their voice services to Voice over Internet Protocol (VoIP), why is LUS — which has no investments in copper infrastructure or switching capacity to protect — considering investing millions of dollars in switching technology? That technology only has relevance in systems burdened by legacy infrastructure. LUS has no such legacy technologies in place, yet the draft feasibility study has them considering equipment purchases that would make sense only if they did.

At one point during the visit in the trailer, I told Pat Sims that it looked to me that LUS was contemplating retrofitting essentially obsolete technologies onto a fiber system in order to make the transition more palatable for customers. He said that's exactly what they appear to be doing. He based this on the fact that the technology deployed in the demonstration box that LUS is including in its public show and tells was actually specified by BellSouth, SBC and the other Regional Bell Operating Companies (RBOCs). This equipment was specifically designed to help maximize existing RBOC investments in plain old telephone service (POTS) and legacy (i.e., obsolete technologies) such as asynchronous transfer mode (ATM) switching technologies.

None of this is necessary (nor, some would argue, even desirable) in a modern, fiber-based system being built from the ground up. None of it is present in the iProvo model touted by Mayor Lewis Billings in his speech at The IndExpo last week. iProvo's system is going to deliver 100 megabits of data capacity to every customer location in that city. The approach backed by the RBOCs (and, it appears, by LUS) will deliver much less bandwidth to customers.

Pat Sims said yesterday that the thinking behind this RBOC-oriented approach is that "customers will not need more" than the 24 megabits or so of data that this approach affords customers.

This strikes me a just flat wrong.

This is a clear example of the RBOC mindset of managing bandwidth scarcity. The purpose of moving to fiber optic systems is to enable customers to operate within the context of bandwidth abundance.

The notion that an RBOC or LUS "knows" with any degree of certainty how much bandwidth customers will "need" flies in the face of the fact that bandwidth usage continues to grow exponentially. Driving that growth is a steady stream of new applications, none of which existed a few years ago. This stream shows no sign of abating. If the experience of places like South Korea, where fiber is being deployed to every home and business, is any indication, ubiquitous access to fiber speeds the growth of bandwidth demand even faster. Erring on the low side of bandwidth provision is what RBOCs have consistently done — much to the chagrin of customers. For LUS to adopt this same approach when building a spanking new network would be unforgivable.

Recognizing that the feasibility study was a draft — a work in progress — John and I had agreed that it would be best to refrain from expressing our concerns pending either clarification from LUS or our gaining access to better information.

Thanks to the fact that ADC brought their trailer to town and we had the opportunity to benefit from our discussions with Pat Sims, I now believe that reservations about what appears to be the overly conservative LUS approach to be this network investment (from a technological, if not a financial standpoint) are more pronounced — and possibly more substantive — than the were before we spent time in the ADC trailer.

My questions have nothing to do with the advisability of the fiber project itself, but rather what the architecture of the network will be. The decisions made now are going to impact the development and evolution of that network — and this city — for decades to come. We should not undershoot what is possible and certainly should not rush to embrace an approach that is based on the concept of limiting the utilization of the network assets. Having taken the audacious step of announcing its intent to pursue a fiber to the premises strategy, for LUS to lose its nerve now and deploy a system that is only incrementally better than systems currently in place would be a tragedy.

I make no bones about it. I view our role here to play the role of maximalists — to point to what is possible and to make the case for pursuing that goal in a vigorous, but realistic manner. Looks like we are going to have to assume that role in earnest now.

No comments: