IPTV Column: The last mile, fiber or copper?

By Daniel Marcus, Product Marketing Manger, Broadband UTStarcom, Inc.

The race to build the telecommunications network infrastructure capable of delivering triple- and quadruple-play services is on. The major players in the U.S. have come to market with trial and commercial offerings for IPTV. These offerings are being developed across a range of different access infrastructure models. The question is: which of these models represents the best choice given cost-constraints and a highly competitive, evolving market for bandwidth-hungry applications?

Not long ago, the debate was framed in terms of high-speed DSL variants versus fiber.  However, rapid adoption of new IPTV interactive applications in the marketplace is driving carriers to shift consideration.  Fiber versus DSL is no longer the framework for the debate.  The  question for carriers now is how far to extend the fiber access network and when to hand off to DSL, if at all.  

In recent months, separate industry and financial analyst firms have publicly stated that both AT&T and Verizon should rethink their deployment strategies for delivering video services. On the one hand, analysts have argued that the cost of a Fiber-to-the-Premise (FTTP) deployment by Verizon, between $600 and $700 per subscriber for Broadband Passive Optical Network (BPON), is too capital intensive and will slow the mass deployment of triple-play services. On the other hand, the Fiber-to-the-Node (FTTN) architecture selected by AT&T will not deliver sufficient bandwidth to accommodate the interactive multimedia and High Definition services required to compete with rival Multi-Service Operator (MSO) and satellite networks. The final outcome is likely to be a combination of both positions.

The one thing that is clear is that wireline telecommunications carriers need to move quickly if they hope to win the battle for the end user. The MSOs are not sitting still while carriers are building out their networks. Bandwidth is not an issue for them. For the most part, the modern MSO network is capable of supporting data rates equivalent to those supplied via a GPON (Gigabit Passive Optical Network) network by minor upgrades and node splitting. Combined with the emerging DOCSIS 3.0 standard, the promise of data rates as high as 100Mbps per subscriber household, becomes feasible, though certainly costly.

Different Architectures for New Business Models

In addition to the competition they face from MSOs and satellite, carriers need to be aware that the business models for delivering video are changing rapidly. Network-agnostic service providers like Akimbo, Skype, Vonage, and Sling Media are threatening both the MSOs and telecommunications carriers with relegation “fat dumb pipe” status by piggybacking their services on the carriers’ access networks. This is a worst-case scenario in light of the significant and costly network upgrades carriers are undergoing to achieve “future-proof” access speeds. In some cases, carriers like AT&T have elected to work with these network-agnostic service providers (in this case Akimbo), in hopes of coming to market more rapidly with a more differentiated offering.

Bearing all of this in mind, carriers need to balance the speed of service delivery, the evolving service package required, and the network costs to find the path to a truly competitive offering. In order to evaluate the best path for carriers to take, it’s informative to review the advantages and disadvantages of these sometimes rival and other times complementary technologies. This serves to better inform other carriers worldwide that are moving forward with their own IPTV strategies.

Each of these architectures has its champions in the U.S. Verizon is committed to delivering service using a PON FTTP network. SBC is focusing on an FTTN model, supplying a fiber connection to existing copper access networks, which are upgraded to support higher speed VDSL in the final segment to the customer premise. SBC has elected to only provide FTTP (Verizon’s model) in Greenfield applications.

PON deployments in the U.S. have primarily relied on BPON, with Verizon leading the way in its deployments of the FIOS system. Whereas a number of municipalities and smaller Telcos have deployed various alternative PON technologies, Verizon’s over 500k lines of BPON represent the clear majority of PON deployments in the US. For Verizon, the delivery of broadcast video is achieved via a separate wavelength overlay, which carries separate analog video signals. In many respects, this architecture is similar to that employed by cable companies in their own digital offerings, and allows Verizon to transition very quickly to the delivery of broadcast services. At the same time, Verizon is able to utilize the high-speed digital nature of BPON to offer bandwidth-intensive interactive services. Verizon is starting to leverage this architecture and is conducting a limited roll out of interactive features, including instant subscriber-customized traffic and weather reports displayed on the TV.

In Asia, carriers overwhelmingly adopted Gigabit Ethernet Passive Optical Network (GEPON). This technology is advantageous in newer networks that do not require support for legacy TDM and ATM services. By contrast, carriers in the U.S. must continue to service their existing customer base with these legacy offerings. GEPON remains a good option for municipalities and other entities that do not require support for legacy services and want to roll out a more cost-effective network that is easily manageable using well-known Ethernet technologies.

Long-term deployment plans for PON in the U.S. and Europe are likely to include a transition path to GPON, as the price for this technology decreases over time.  With GPON, carriers will ultimately be able to deliver up to 100Mbps per subscriber, enabling an evolving suite of interactive services.

Weighing the Costs to Extend Fiber to the Consumer

Given the physics of delivering DSL service, carriers have had to redefine the concept of the carrier serving area. For VDSL services, the loop lengths need to be significantly shorter than the 5kft number that AT&T is reported to be using in its projections. To deliver 50Mbps rates, the service needs to be deployed within 1000 feet of the subscriber, thereby driving up the cost for the fiber required in the network. In addition to loop lengths, the quality of the copper is a significant contributing factor to achieving optimal speeds.   Relative to MSO architecture, which requires reconfiguration, potentially considerable upgrades, and integration, new FTTN deployments support the full gamut of interactive IPTV services from the onset, thereby allowing a much richer feature set.

Unquestionably, carrier costs rise commensurate with their projected fiber outlay, which carriers must weigh against their near-term uptake and ARPU projections. The largest costs for delivering PON result from the final 100 yards of installation to the customer. Frequently, in installation environments wherein the final connection to the customer must be buried, it’s common for the carrier not only to pay for trenching the fiber, but to also incur the costs of repair to driveways and lawns. This can impact time to service roll out, both in terms of the actual work required and the time necessary to receive permission by a variety of local communities, municipalities, and home owners.

Cost for Customer Premise Equipment (CPE) is also considerable. While this is expected to decrease in the future, for now the cost of CPE for FTTP applications, at $350 per household, is at least double the costs of deploying FTTN with DSL in the final segment. GPON is the ultimate enabler for the 100 Mbps forecast attributed to most U.S. Fiber-to-the-Home (FTTH) plans. At this time, multiple vendors have announced products but so far only a couple claim to be shipping; testing has been inconclusive, and none are deployed in commercial networks. By contrast, DSL CPE is considerably cheaper than their optical counterparts (ONUs) and there are a significant number of established vendors shipping stable units, driving costs even lower.

Time to market for an FTTN deployment is better due to the fact that the carrier can leverage its existing copper plant, thereby eliminating the time required to connect the customer to the service. Upgraded DSLAMs and their respective CPE are required in order to realize the higher speeds of the latest DSL/VDSL variants. But again, these are significantly lower costs than their FTTP counterparts.

In addition to the equipment and time cost disparities, the costs of deploying FTTN can rise steeply as the density of the homes served decreases. In a heavily populated area, the carrier can reach a large number of customers and deploy the DSLAM over short loop lengths, which allow the required higher speeds that are predicted for future services. This has long been the case in Japan and Korea where telecom carriers routinely offer access speeds in excess of 20 or 30Mbps.

The cost of FTTN is ultimately dependent on the take rate of the service. As more customers sign up, the fixed-cost portion of the service is leveraged across a larger subscriber base and the average cost per subscriber decreases. At the onset, the cost per subscriber is much higher and this is compounded in the West, with the higher cost of the newer GPON technology. According to a recent Cowen and Company report, Verizon’s FIOS service costs an estimated “$1,605 per connected home, but [is] expected to fall to $677 by 2009.” (Cowen and Company, May 25, 2006).

In conclusion, both FTTP and FTTN approaches will likely find appropriate venues.

FTTN allows carriers to move forward with their IPTV deployments more quickly and it can be cost effective, particularly in densely populated areas and MDU applications.

Long term, FTTP will afford the biggest advantage to carriers. In the short-term expect to see many hybrid solutions.