Glut or No Glut?
Graham Finnie, 2-Jul-2001
The argument about the bandwidth "glut" in Europe and elsewhere has taken on a near-religious character in recent weeks. Some say there's far too much, while others counter hotly that there's too little.
This passion is hardly surprising: for wholesale bandwidth providers, fibre-optic network construction companies, optical equipment vendors, ISPs and many others, this is a life-and-death issue. At root, there are three reasons why analysts are so divided on this issue: the first two favour those who believe there's no glut, but the last may hold the key to the real issues that confront the industry.
The first big gap, between lit and latent capacity in fibre-optic networks, quickly confuses those who haven't fully taken this into account when analysing how much capacity is unused. Before dense wavelength-division multiplexing (DWDM) became available, all the bandwidth in a cable was generally available from the day it was ready for service. Today, however, very little is in play from the start.
In transatlantic cables, for example, a new TeleGeography report shows that only a small portion of each cable is equipped with DWDM lasers and detectors on the day it comes into service. In 20 long-haul submarine cables that came into service over the past 18 months, less than 20% of total available capacity was lit initially. And the gap could widen as even higher-capacity cables with higher wavelength counts come into play.
The second big bandwidth gap that leads to misunderstandings is the gap between what gets lit and what consumers and businesses actually use. The most important reason for this gap is the recent emergence of a wholesale telecoms segment, creating one or more layers between the network construction companies and the users. More capacity is usually purchased at each layer than is actually required, and the gap between availability and demand is growing for a variety of reasons.
For example, as prices fall, buyers have been tempted to widen the gap between deployed bandwidth and actual demand in order to improve quality (especially in IP networks) or guard against unexpected demand spikes and shortfalls. Since the time it takes to provision new bandwidth or circuits is actually increasing in some cases, this is often a sensible move. From the point of view of the network builders, however, a sale is a sale-even if the unsold capacity isn't used.
Finally, and most controversially, there is the semantic and literal gap between access and long-distance networks. When some analysts speak of a bandwidth dearth, they are really talking about the lack of capacity in the access network, which is a bigger problem in Europe than in the US.
There's near-universal agreement that broadband networks will give a massive boost to demand and, if widely deployed, could easily soak up the capacity that DWDM is bringing into long-distance networks-but there's disagreement about when that will happen. Broadband has barely begun to make an impact in Europe, except in Germany.
That, however, raises a another issue that is ultimately more important than all the others: even when latent capacity and redundant or spare capacity aren't factored in, it looks like many key routes will have a lot more capacity than the market requires. The huge latent power in new fibre technologies will be of no use if businesses and consumers don't find new applications compelling enough to justify spending more on telecommunications.
And the uncertainty about those applications feeds reluctance to invest in any technology with no obvious short-term payback. The result is that many broadband access build-out plans are being put on ice, and the gap between long-distance and access networks continues to grow. Bridging that gap is the most important single task facing the industry.
Graham Finnie is consultant, TeleGeography (US) and The Yankee Group Europe . He can be reached at graham@finnie.com. |
