Business/Government Broadband Access Market Trends
By, Stephen Montgomery ElectroniCast

CABLE LINK COMPONENTS

The global consumption of selected broadband communication components in Business/Government Access, estimated at $3.6 billion last year (2005) is forecasted to reach $4.9 billion by 2008.  Optical transmitter/receiver units will hold the leading fiber optic component consumption value share and is forecasted to reach $1.3 billion in a couple of years. The fastest growth in fiber optic components used in Business/Government Access usage will be by passive fiber components, multiplying more than 10-times (10X), over the next 5-years.

Cable link components constitute the largest current value share of total Business/Government broadband network component consumption. The deployment of new fiber cable has not made necessary, to date, a matching increase in investment in central office and subscriber equipment.  Local exchange carriers (LECs) and other service providers, after the 2001-2004 caution, began to accelerate their move into Fiber-To-The-Office (FTTO) network deployment.  This initial deployment, however, will provide adequate cable capacity for projected growth over the next 10-15 years, while central office equipment and subscriber equipment will be added only as subscriptions justify, year-by-year.

Business/Government (office access)               The 2006 global deployment of fiber FTTO cable link components (cable plus connectors) is estimated to run about over $800 million.  Fiber cable link component consumption in business/government (office access) networks will grow to $1.3 billion by 2008.  This consumption will be dominated by cable, but connectors will maintain a significant value share to reach over 10 percent share, by 2010.  Access networks will require numerous connector interfaces between the central office and subscriber; to splitters, WDM filters, installation apparatus and other points.

The North American consumption of fiber optic and other broadband components, in broadband-to-the-office interconnection is expanding steadily and is forecasted, by ElectroniCast, to reach $1.7 billion by 2008.  The leading component will remain to be transmitter/receiver pairs, which currently (2006) hold a 36 percent market share.

As soon as Gigabit Ethernet (GbE) began ramping up in deployment, the focus shifted to next-generation 10 GbE (10 gigabits).  Besides supporting super-LANs, 10 GbE is expected to move into the general service provider networks; Telco SONET, Internet and converted cable TV.  The main focus is 10 GbE standards.  Basic to the planning is the assumption that 10 GbE will follow the same access method, management objects, frame format and full-duplex operation as the earlier 10/100/1000 Ethernet.  Development efforts include fiber and copper components, interconnect media and equipment.

At the device and component levels, there is major contention about whether 10 GbE transmitter and receiver elements should use single emitters (serial transmission) at 10 or 12.5 Gbps, or whether the links should be parallel; for emitters.  A secondary contention, from the device vendor standpoint, is whether these emitters will be VCSEL diodes or edge emitter laser diodes. For shorter distance interconnect at 10 GbE, standards development has encompassed twisted pair copper interconnect; an improved copper, category 6, aimed mainly at interconnect within wiring closets at 10 gigabits.  10 GbE is being used mainly for connecting high end switches and routers to the backbone.  The design target has been 20 Km reach of 10 GbE, over singlemode fiber, compared to 5 Km for GbE, and 300 meters over multimode fiber.

Nearly all fiber optic cable now being deployed for broadband local loop applications is optimized for the 1310 nm band.  This is in contrast to the 1550 nm band widely used in long haul and interoffice networks.  Optoelectronics for 1550 nm are substantially higher priced than equivalent performance 1310 nm band optoelectronics. Local exchange carrier planners have determined that, for fiber spans below 50-80 kilometers, the 1310 nm band with associated electronics is the most economical choice, even though the fiber loss is about twice as high.  A few users have chosen 1550 nm, especially for their longer trunks, mainly so that they could use optical amplifiers, which have not been available for the 1310 nm band.  A concern about the earlier gstandardh singlemode fiber has been its attenuation peak in the 1450 nm region, between the commonly used 1310 nm and 1550 nm bands.

Trend to Duct Cable                Nearly all the recent fiber optic cable deployment for FTTO services has been by incumbent local exchange carriers and competitive alternate local carriers (CALCs).  This initial thrust has been concentrated in the higher density metropolitan areas, making substantial use of the existing underground ducts.  As FTTO fiber deployment spreads to the suburbs and rural areas, however, the cable mix will shift to a much larger share of aerial installation cable.  Aerial installation on existing poles is relatively inexpensive, while digging trenches for cable installation is prohibitively expensive.  As new contenders (local exchange carriers, Interexchange carriers/IXC, and others) enter the FTTO race, they too will initially target the high office density areas that are relatively accessible via existing ducts.  Over the forecast period, however, the aerial cable share in North American office links will lag in deployment, shifting from about 13 percent value share last year in 2005 to 11 percent in 2008.  The buried cable share will be relatively minor, over the forecast period.

Utility Rights-of-Way             A major item in the planning and deployment, plus financing, of a fiber network is the acquisition of rights-of-way (ROW).  Public utilities (as well as city and state governments) have established ROWs, and are leveraging this into deploying massive long haul fiber capacity and leasing or selling fibers, and/or wavelengths on fibers. ElectroniCast estimates that, in the United States, electric utilities and municipalities own nearly10% of the installed-based of fiber optics cable.  Nearly all of the large IOUs (Investor owned Utility) or government owned utility has extensive fiber optic cable networks. Since they already own the rights-of-way, the poles and the towers, power utilities have the capability to economically deploy cables parallel to their powerlines.  The top 10 electric IOUs in the United States, based on the number of installed distribution infrastructure (poles and towers) are listed below.

                Top 10 Electric IOUs in the United States

(1)            American Electric Power Co.
(2)           Pacific Gas and Electric Co.
(3)             Duke Power Co.
(4)              Alabama Power Co.
(5)              Texas Utilities Electric Co.
(6)               Pennsylvania Power & Light Co.
(7)               Consumers Power Co.
(8)              
Niagara Mohawk Power Corp.

(9)               PacifiCorp
(10)          Carolina Power & Light Co.

Fiber Optic Connectors              Last year the use of fiber optic connectors in FTTO networks was a relatively modest value ($57 million), compared to the associated cable.  Nevertheless, by 2008, this consumption will reach $102 million.  For every connector attached to a fiber optic cable, there is an associated non-cable optical connector to which it mates.  One significant use category is connectors built into transmit/receive modules.  Another major market consists of connectors built into fiber optic installation apparatus.  About 30-70 percent of the total value of patch panels, distribution frames and other installation apparatus consists of built-in connectors. 

DWDM   The global consumption of dense WDM components, segmented by type in fiber-to-the-office interconnection is estimated to exceed $160 million this year. The leading DWDM component will be passive DWDM filter modules, with a 75 percent value share in 2006.

Dense WDM Components Global Consumption Value

In FTTO Local Loop Deployment, by Type

        

Direct Modulated Lasers for 10 Gbps Access      The direct modulated laser (DML) module has been developed by several vendors, for 10 Gbps interconnect on conventional singlemode fiber, at 1310 nm.  The primary focus is low cost, for intra-city and short-reach interconnect.  Operating in the 1310 nm band, chromatic dispersion (and the related, expensive dispersion-compensating modules) will not be a factor.  DMLs are now in high volume use, OC-12 metro.  OC-48 DMLs also are widely used in 1550 nm and 1310 nm WDM and local access applications.  The DML also can provide substantially higher output power than an externally modulated laser diode, reducing the need for optical amplifiers.  The higher data rate (OC-48, up) DMLs generally incorporate distributed feedback (DFB).  DML laser transmitters will be used extensively in SONET/SDH rings, including optical add/drop multiplexers.

Rapid Growth of DWDM         Wavelength division multiplexing has not been used significantly, to date, in business/government broadband access networks.  However, as these networks evolve and as per-channel optoelectronic costs drop, DWDM will gain rapid acceptance in these networks.  Fixed configuration passive DWDM filters will monopolize the filter function in FTTO networks, with an estimated value share of over 95 percent; in 2006 (These values exclude WDM filter functions monolithically integrated into multichannel transmitter/receiver modules). There will also be substantial use of non-dense (gcoarseh) WDM filters in business access markets. 

Planar Waveguide Amplifier for Access        A promising alternative to pumped doped fiber amplification is amplification in planar waveguide, as shown below.  Amplifiers based on this technique are commercially available for operation in the 1520 nm band, and have been demonstrated in the 1310 nm band.  The incoming light beam, with signal, is coupled from the singlemode fiber into the planar waveguide, amplified by pump, and then coupled back into the ongoing fiber.  The planar waveguide is much more amenable to a wide range of dopants, concentrations and dimensions, compared to the fiber.  This increases the feasibility of developing amplifiers for wavelengths not yet in use.

Access Network Planar Waveguide Optical Amplifier














Because 1310 nm and 1550 nm signals can be transmitted down the same fiber simultaneously, other interesting options are available.  One example is a dual-window DWDM, combining DWDM of both bands.  The legacy singlemode fiber at 1310 nm, can support 10 Gbps in each DWDM channel, and higher, without dispersion compensation.  Conversely, the S band channels must be separated out of the fiber from the C band channels before dispersion compensation is applied to the 1550 nm band channels.

A general trend in fiber optic components is to greater integration; combining more of the individual components in a functional package, into a single, continuous-waveguide functions.  This is proceeding now, in two generations:

Gen 1 fused fiber interconnect

Gen 2 monolithically processed planar waveguide interconnect in substrate

Many Factors Driving Integration      Dense WDM transmitter and receiver systems are high-value combinations of active and passive components.  Increased density, cost reduction, performance improvement and competitive advantages all contribute to the movement toward integration of these elements. Concerns about lifetimes of active units (transmitters and receivers), and the relatively low production lots by individual vendors, supported piece-part assembly, rather than integration, in early production, 1995-2002.  Current volumes, however, along with the trend to many more elements in a typical module make integration an attractive alternative. Density is valuable.  An integrated 32-wavelength DWDM module requires only a few percent of the volume required for a module assembled from individually packaged components. For singlemode components, assembly and test labor is a substantial part of the total cost.  If volume is sufficient, this cost (including factory overhead) can be substantially reduced by automated assembly and test.  Improved reliability is an added bonus.

The transmitter DWDM module interconnects a number of precise-wavelength transmitters to fibers of a passive, waveguide combiner, feeding all wavelengths into one singlemode fiber.  There is a trend toward incorporating wavelength lockers at the output of each transmitter.  Other functions, such as optical monitor taps and a 1xN switch with a wavelength-tunable backup transmitter will be later additions.  The receiver DWDM module combines a DWDM de-multiplexer filter with the required number of receivers.  Remotely tunable attenuators are a possible future addition, along with a wavelength-switching matrix and/or output wavelength converters.

CWDM is less expensive than installing a second cable         The 1310/1520nm CWDM is being popularized by access passive optical network (PON) deployment.  This will be used mainly in residential fiber deployment.  However, there will also be significant PON deployment to offices, particularly by ILECs and CLECs. Transport service providers will choose the lowest-cost solution.  Initially, in ggreen fieldh situations where loop fiber must be deployed to subscribers, lowest cost generally is to install more fiber, versus less fiber plus CDWM. Most fiber installation consists of installation labor plus right-of-way costs, relatively independent of number of fibers in the cable.  Later, however, when the installed fibers approach total use, the carrier often will find that CWDM is less expensive than installing a second cable.

CWDM components are significantly less expensive than DWDM components.  Unit price of both active and passive components is inversely related to wavelength spacings.  The design and fabrication of DWDM components is technically challenging (and increasingly so, as spacing shrink), but CWDM components are easy entry for all fiber optic component vendors.