Metro Ethernet, Optical Access: Still Far to Go

In the enterprise high-capacity access markets, one has to distinguish between the financial and operating markets. Of late there has been renewed interest in the financial value of scarce optical assets, particularly in smaller markets.

But the allocation of new capital to the access business, if welcome, is not the same thing as deployment of capital to support alternate optical access facilities to the places most businesses are located, which is, simply, in the larger markets.

There is no “silver bullet” in the optical access market; just determined, steady, slow progress in lighting new buildings with at least one fiber cable. To be sure, global carriers very much want to connect large enterprise locations with 1 Gigabit-per-second to 10 Gbps optical connections.

The problem sometimes is that such connections don’t exist, or sometimes simply that sourcing such facilities is laborious because there are so many small providers in local markets. The problem for a global carrier is simply the need to source really high bandwidth access all over the world, easily.

In part, it’s a Layer One issue. In the U.S. market, for example, only 12 percent of business sites have fiber connectivity. Only 20 percent of North American cell sites have fiber connectivity.

That explains the continuing attraction wireless and Ethernet-over-copper alternatives represent. To be sure, programs such as Verizon’s FiOS will solve those problems for consumers, and almost incidentally for many branch office, small office or smaller business executives.

In the second quarter, for example, Cogent Communications added 1,208 on-net connections, up 53.5 percent from the 787 added in the first quarter. In the third quarter Cogent added 30 buildings and expects to have added 100 on-net buildings by the end of the year.

The company expects to do so again in 2008, adding 100 new buildings to its network.

“As of September 30, 2007, we had 1,189 buildings directly connected to the network, representing over 520 million square feet of rentable office space, out of an addressable inventory in North America of about 6.2 billion square feet,” says Dave Schaeffer, Cogent CEO.

“We are currently utilizing a little bit less than 22 percent of the lit capacity in our network,” says Schaeffer, illustrating the issue nicely: fiber isn’t the problem, access to customers with fiber is an issue.

At the end of June Time Warner Telecom had 7,884 buildings connected on its own facilities. At the end of September the company had 8,109 buildings on network, an increase of about 225 buildings, or about three percent. On an annual basis, on-network buildings increased about 19 percent.

RCN has something in excess of 800 buildings on network. Optimum Lightpath has about 2,500 buildings on network with fiber connections.

Nationwide, there are some 95,000 fiber-fed buildings, says GeoResults. And of course, compounding the problem is the fact that lots of the fiber access to lit buildings is in a common cable sheath, no matter who the retailer of record is. For many desirable buildings, the issue is that most of the suppliers actually use fiber in the same cable sheath.

There is progress. It simply is progress of the persistent, gradual sort.

The point is to separate the legitimate financial plays—rolling up and aggregating optical access assets in tertiary markets, such as Zayo Bandwidth is doing, from the operating situation, which continues to be that optical connections to more buildings is the gate.

One would think optical connections to wireless towers are an obvious, slam dunk sort of opportunity. With broadband demands growing rapidly, and locations so easy to identify, replacement of copper-fed T1 or microwave connections, the typical solution these days, would seem to be a fairly easy business proposition.

There are perhaps 2.2 million wireless base station sites globally, including 250,000 in North America alone. Assume half those base stations use wireless backhaul, while the other half use leased T1s or optical connections.

The Chinese market is unusual in the sense that most of China Mobile’s base stations already are fiber connected. Observers tend to note that in Europe, the Middle East or African markets, it wouldn’t be unusual to find that 60 percent of connections use microwave technology while 25 percent use optical connections and just 15 percent or so are based on copper E1 connections.

In the U.S. market, perhaps 10 percent to 20 percent of towers and other transmitting locations use fiber connections, accounting for 25,000 to 50,000 optical backhaul locations. And though microwave backhaul is popular in other markets, it rarely is used in the U.S. market.

That suggests as many as 225,000 wireless tower sites, or as few as 200,000, are fed by T1 connections over copper media. Depending on which carrier is involved, backhaul can represent 20 to 40 percent of recurring operating cost.

Verizon and at&t obviously are in position to use their other assets to slice this cost of doing business, while Sprint Nextel and T-Mobile obviously face higher costs. But the fiber access opportunity isn’t necessarily contingent on replacement of copper-fed T1s with optical replacements.

Indeed, voice works pretty well when the backhaul is based on T1 technology, so carriers might well not want to complicate their operations by moving all that traffic over to optical access. It might in fact make just as much sense, or more sense, to use the optical facilities for the rapidly-growing IP traffic demands, leaving T1 facilities in place for voice.

In other words, use the Time Division Multiplex network for voice traffic that is highly sensitive to latency, and use optical Ethernet for bursty data traffic. Of course, thinking is bound to change once any appreciable amount of usage and revenue is generated by video.

At some point, optical will be the best choice. The issue is when that will happen, and what the optimal choices are in the meantime. The point is that optical Ethernet, though the long-term answer, doesn’t cleanly address all the operational issues carriers think they face.

Encapsulating TDM traffic for Ethernet transmission worries carrier technologists for any number of reasons, for example.

The bottom line is that optical Ethernet, and business optical access, continues to grow every quarter. It just isn’t the sort of transformation that can happen much faster, given the need to balance revenue from the first customer account with the cost to construct an optical lateral connecting that customer.

In the old days, when carriers were the primary customers, matters were simpler. One simply built out to carrier hotels, data centers and key central offices, knowing that most of the high-bandwidth termination demand would be at such locations. That isn’t so easy when the customer base primarily is enterprise customers.