Fiber-to-the-home--Why not now?

FTTH networks are pricey and have some operational issues that must be addressed before they become widely used.

By Don Gall and Mitch Shapiro

Almost from the conception of fiber-optic cable, fiber-to-the-home (FTTH) -or the passive optical network (PON)-has been envisioned as the network of the future. Singlemode fiber combines very low signal loss per kilometer with high carrying capacity. On the surface, these traits should make an ideal network with very high bandwidth availability, ultra reliability, and the ultimate in network futureproofing.

The relatively recent development of ITU grid laser transmitters, broad-bandwidth EDFAs, extended operating windows in fiber-optic cable, and dense wavelength-division multiplexing (DWDM) technology have made this type of network technically feasible.

FTTH trials have been underway for well over a decade. The first of these used multimode fiber, which was drastically bandwidth-limited by dispersion and had a very high signal loss per kilometer. Even very limited local area networks (LANs) using this technology were much more expensive then competing technologies.

The advent of singlemode fiber and 1,310-nm laser diodes eased the bandwidth and distance limitations but was still extremely fiber intensive (expensive) for fiber-to-the-curb (FTTC) applications. Next in the timeline were 1,550-nm lasers and simple WDM techniques that cut the overall fiber counts in half by allowing two frequencies on the same fiber. More FTTH trials were launched and both the telephone and cable industries started using fiber optics to enhance their existing networks. But true FTTH was deemed to still be a somewhat distant dream.

A recent set of lab tests by Bell Labs suggests that more than 1,000 frequency-stable laser transmitters can be multiplexed onto a single fiber with up to 160 Gbits/sec of data on each wavelength. That suggests the new upper bandwidth limit for a single fiber is approximately 160 Tbits/sec. That's enough capacity to have every person in the human race talking to each other on one fiber (2,500,000,000 calls at 64 kbits/sec)! Today, common practice is closer to 10 Gbits/sec per wavelength, using eight ITU grid lasers, spaced on 200-GHz centers in the 1,550-nm window.

Although eight lengths are nowhere near the potential capacity of a single fiber, it is very close to the bandwidth necessary to make FTTH feasible. Today, singlemode-fiber cable is relatively inexpensive to deploy. Fiber cable designed to use the 1,310-/1,550-nm windows like Corning's SMF-128-type fiber can be purchased in quantities for less than 3 cents per fiber foot. In most applications, fiber can be installed on a level playing field with more traditional networks (twisted-pair and coaxial cable), without even factoring in the tremendous upside in capacity.

Distribution network electronics are readily available from several sources. The headend/hub/central-office equipment is a shared resource that is very cost-effective when compared to competing technology. It is the network interface to the residence that becomes expensive versus other alternatives. The device(s) must terminate the fiber and convert the optical signals to an assortment of services such as telephone, high-speed data, meter reading, energy management, security, and television service.

Today's solutions are all over the map, from engineering prototypes to second-generation products. The least expensive solution that does an adequate job of covering services is priced at more than $1,000 per household in large quantities. This is significant since DSL and broadband solutions are considerably less expensive. To compete, the home- network interface will need to be integrated into a much less costly package.

On a pure cost basis, DWDM components are also a major factor in being competitive with other network alternatives. Using current pricing, the cost of a pair of 8-port DWDM modules runs from $700 to $1,000 per port.

These prices work in the transport segment of networks where a single DWDM pair can yield an eight-fold increase in capacity. DWDM does not make sense today in distribution, however, where it could conceivably be used to feed eight residential customers.

We feel that the price per port needs to drop by at least a factor of five to be competitive. We also believe that EDFAs will need to become flatter across their frequency range, and DWDMs will need more consistent insertion-loss specifications in any application that involves analog optical signals.

Network powering is another issue that needs to be addressed. Although it has been proposed jokingly by a few optical humorists, it is going to be impractical for a very long time to power any network electronics by using a laser source. This situation leaves us with a choice between powering from the home or building a power-distribution network that parallels the fiber distribution.

Powering from the home becomes problematic if the power source needs to be backed up to meet lifeline telephony requirements. Current backup options use standby batteries that at best have only a four- or five-year lifecycle. Imagine having to be responsible for replacing and arranging for the disposal of thousands of batteries every year.

There may be a better answer in the future using very small, inexpensive fuel cells. The other alternative is to use twisted-pair or coaxial cable to provide the power using a centralized source. This solution is technically practical but very expensive.

Last but certainly not least is the practical side of distributing fiber to a vast number of sources and keeping the mean-time-to-repair statistics within a respectable range that is acceptable to your customers.

Singlemode-fiber cable has been designed for and deployed almost ex clusively in the transport segment of most networks. Today, it is packaged on cable reels in multiple kilometer lengths to reach between splice points placed as far apart as possible.

This approach is a far cry from the suburban distribution network, where the idea is to drop one or more fibers off at each residence. To properly facilitate FTTH, there needs to be a solution that will be flexible enough to distribute fiber anywhere from zero to eight residences at each pedestal or telephone pole with a minimum of splicing labor and fiber waste. In an average residential area in the United States, this distance averages roughly every 150 feet. Unless there are some very large leaps in the overall reliability of mechanical connectors, this portion of the network should all be fusion spliced.

Then there are maintenance issues. In singlemode fiber, the area that carries tha optical signal is only nine millionths of a meter across. A speck of dirt you can't see with the naked eye can totally block the optical signal!

Our experience with fiber suggests that the cable itself is as robust and reliable as any of its competition. Properly fused splices are time-consuming and have to be very technically precise (compared to splicing coaxial cable and twisted-pairs) but once completed are also very robust.

Our concern is the new environment for fiber in a FTTH network-one full of destructive pets and people with sharp objects (shovels, garden tillers, pinking shears, etc.). It is also full of fences and landscaped lots that are not always easy to access at night or in the middle of a storm.

To complete a fusion splice, you need a relatively dry and clean place with limited interference and a $6,000 to $45,000 fusion splicer. Twisted-pair and coaxial cables can be spliced with parts that cost pennies and hand tools that cost less than $100. Customers will love the reliability and network capacity, but they will not tolerate service outages that last two or three times longer than competing technologies. To really get serious about FTTH, we need a better solution for maintenance than exists today.

Don Gall has been involved with the cable-TV industry for the last 28 years. He was an integral part of the team at Time Warner that developed the first practical applications of analog fiber and hfc networks. He is currently a consultant with Pangrac & Associates (Port Aransas, TX) and can be reached at [email protected]

Mitch Shapiro has been tracking and analyzing the broadband industry for more than 12 years. He is currently a consultant with Pangrac & Associates, which this summer will publish the first of a series of in-depth reports on clustering, network upgrades, and new service strategies in the cable industry. He can be reached at [email protected] or via the P&A Website at http://broadbandfuture.com.