Editor's Note: Due to a combination of unreviewed edits and other factors, the "Analyst Corner" piece from David Dunphy of Current Analysis published in the November issue did not accurately reflect the company's viewpoint on the DSL and PON markets. Mr. Dunphy has provided a corrected and updated article, which appears here in its place. The article will also appear in our January 2005 print edition. We regret any confusion the November issue may have caused.
SBC chose Alcatel's IP-DSL access multiplexers (DSLAMs) as its principal means of building-out broadband access to its installed base of residential customers, giving Alcatel's 7340 fiber to the premises (FTTP) equipment a secondary role. Verizon, on the other hand, chose AFC, selecting a broadband PON (BPON) approach to make an optical FTTP solution its primary means of building-out broadband access. These service operators had collaborated (along with BellSouth) on developing a set of requirements that went into a joint RFP, which might make it seem surprising that they did not select the same technology. Did Verizon or SBC disagree about residential bandwidth requirements or the reliability or total cost of owning DSL or FTTP? Not necessarily. The reason that SBC and Verizon chose different technologies for broadband rollout had more to do with their existing networks than with differences between FTTP and DSL capabilities.
FTTP offers greater scalability than DSL; ITU-T G.983-based BPON systems support 622 Mbits/sec downstream and 155 Mbits/sec upstream, and ITU-T G.984-based gigabit PON (GPON) systems offer 1.2 Gbits/sec downstream and 622 Mbits/sec upstream. In contrast, ADSL offers cost advantages but lacks similar scalability. Yet, while ADSL in practice typically supports only around 512 kbits/sec to 1.5 Mbits/sec between the DSLAM and DSL modem (in theory it can support 8 Mbits/sec downstream and 800 kbits/sec upstream at up to 18,000 ft/5,460 m), enhancements in next generation DSL technology that DSL vendors worked on while PON vendors proved in their technology and drove down cost to a more competitive level increased the competitiveness of DSL significantly—making it a stronger option for the broadband buildout.
ADSL2, defined by ITU-T G.992.3 and G.992.4, adds 600 ft (about 180 m) of distance over ADSL's capabilities and can transmit 10 Mbits/sec reliably up to 10,000 ft (just over 3,000 m). ADSL2+ (ITU-T G.992.5) increases downstream rates to as much as 24 Mbits/sec for up to 3,000 ft (or 20 Mbits/sec for up to 5,000 ft/1,500 m), and very-high-bit-rate DSL (VDSL) supports data rates up to 52 Mbits/sec over distances to about 4,000 ft/1,200 m, covering 70% or so of the distance requirements in the Europe/Middle East/Africa (EMEA) region and a smaller percentage in North America, where loop lengths are longer.
While DSL solutions themselves cannot replicate the distance or bandwidth capabilities of PON, DSL can be brought closer to customers in a hybrid fiber/copper rollout where fiber provides backhaul for DSL. And DSL supports the bandwidth and distance needed by most subscribers or applications in use today as well as those applications for which strong demand seems probable over the next product lifecycle. Streaming high-definition television (HDTV), for instance, requires about 19 Mbits/sec uncompressed (and compression technologies may reduce that to 8 or 9 Mbits/sec). ADSL2+ can provide one HDTV channel, 4 Mbits/sec of Internet access, and voice. VDSL can support two HDTV channels, 10 Mbits/sec of Internet access, and voice, which for subscribers within the 1,200-m loop limit should provide more than enough for residential and even many small-business applications.
Verizon's greater installed base of aerial fiber (Verizon has been quoted as indicating over 55% of its fiber is aerial) means it faces less trenching cost to deploy FTTP than SBC, which has less aerial fiber in its outside plant. Verizon has estimated aerial fiber is 30% cheaper to deploy than burying fiber and therefore probably has a lower cost structure for an FTTP solution than does SBC. FTTP solutions may still have a higher cost per premises than DSL, but FTTP also offers advantages in future proof scalability that must be considered. Thus, relatively lower cost in installing FTTP at Verizon due to the large installed base of aerial fiber, in conjunction with the benefits of greater future bandwidth scalability, may well have resulted in the win for FTTP at Verizon—while the higher costs of installing FTTP at SBC may have relegated it exclusively to new builds for now. Both SBC and Verizon were testing aerial as well as buried cable, but the impact of the business case for FTTP was greater on the vendor that had more 'glass in the sky'.
Neither Verizon nor SBC are likely to implement only one technology. SBC has already indicated it is supplementing DSL with BPON, and Verizon is likely to supplement BPON with DSL in some locations where that may be more feasible. Though vendors fighting over market mindshare tend to position for "all or nothing" wins, as is often the case, both DSL and PON technologies will play some role in the networks of these two RBOCs.
And though installed base is always an important consideration, things can change over time. Given the installed base of ATM equipment and the greater experience BPON solutions had garnered to date, it was to be expected that BPON would win for now with the conservative RBOCs. But with the strong emphasis on new services delivered over Ethernet, GPON will become increasingly attractive. GPON vendors in the future will have installed base issues to overcome, and it is always difficult to unseat an incumbent vendor. Given the much larger-scale deployment of BPON that can be expected at Verizon, in the future it will be a much harder customer to crack with GPON—and SBC will present GPON vendors with a potentially better target, since it will have less vested interest in BPON technology.
David Dunphy is principal analyst, optical infrastructure, at Current Analysis (Sterling, VA). He can be reached via the company's Website, www.currentanalysis.com.