DSL (Digital Subscriber Line) update
techweb.cmp.com
Smoothing The Bumps On The Megabit Highwat
By Saroja Girishankar ÿ At a revolutionary time for data networking, the "last mile"--the infrastructure leading to businesses and homes--remains the final impediment. Now, with technologies such as DSL (Digital Subscriber Line), even that potholed pathway is quickly moving to superhighway status. In fact, telephony is being turned on its head as slow-growing voice networks (4 percent annually by some estimates) ride the crest of data networks designed for business apps, remote access, the Internet and electronic commerce.
The promise of DSL, and its asymmetric, symmetric, high-bit-rate and very-high-bit-rate (ADSL, SDSL, HDSL and VDSL) variants, lies in its ability to split the frequencies in a single copper loop to handle simultaneous voice, data and video transmission at speeds of 7 Mbps to 9 Mbps.
TeleChoice consultant Kieran Taylor is optimistic that by 1999, volume deployment of DSL will drive down monthly line costs to $50--a pittance compared with today's $800 to $1,000 for T1 services. More conservative estimates put the 1999 monthly line charge nearer $500 to $800. Taylor, however, says 20 to 30 CLECs (competitive local exchange carriers) are offering DSL services at speeds similar to T1 at about half the price of T1.
Although most incumbent local exchange companies lag their CLEC counterparts, U S West was expected to become the first regional Bell to make DSL commercially available. Its new MegaBit Services are slated to be available this month in Phoenix at 192 Kbps ($45 per month), 320 Kbps ($65) and 704 Kbps ($130); unlimited Internet access will be available for about $20 more. In addition, corporations and ISPs can purchase the services for traffic aggregation and oversubscribe the 704-Kbps MegaBusiness service up to DS-3.
The strategy puts U S West ahead of the regional Bell pack. Unless directly challenged by high-speed cable modems or CLECs offering DSL, these incumbent telcos have an incentive to drag their feet--primarily because DSL can cannibalize some higher-priced T1 services. Of course, DSL and T1 aren't totally interchangeable. Taylor points out that T1 can support more voice channels than DSL and, because T1 is symmetric, it doesn't have the upstream bandwidth limitations of DSL (600 Kbps to 1 Mbps). Users also have more T1 equipment choices today and greater confidence in T1 reliability. DSL, on the other hand, is cheaper and pivotally positioned for high-speed Internet access, especially multimedia e-commerce tasks.
For now, though, DSL is generally considered the quickest and cheapest route to high-speed local access for business. And among its variants, HDSL has the most to offer immediately, because of the stability it has gained through existing deployment in conjunction with T1.
DSL's potential has catalyzed interest among service providers. Alcatel says it has recently designed and installed equipment for Ameritech, BellSouth, Pacific Bell, Southwestern Bell, Singapore Telecom and Telia in Sweden; Paradyne claims to have more than 100 business partners.
Others jumping on the DSL bandwagon include Bell Atlantic and UUNET, which already offers services and plans to support more than 200 points of presence by year's end. Both GTE and MCI have promised to begin nationwide DSL rollouts in 1998.
Should businesses jump? Perhaps not just yet. The full potential of DSL won't be achievable until late 1998 or early 1999, when large-scale deployment is forecast. Standards, interoperability and economic hurdles must also be leveled for this still-nascent technology to live up to its potential. Now is the time, however, for businesses to pay close attention to DSL when building or buying platforms and applications.
What Lies Ahead Businesses also may want to begin discussions with providers about their needs--especially because most providers are still building business models for their DSL offerings. Key issues--like how LAN, Internet and intranet access will be bundled--must be resolved. WorldCom, for example, is bundling local, long-distance, Internet and LAN access services with DSL.
Carriers also must address issues of efficient provisioning, billing, security and management of these services to avoid the kind of public relations nightmare that accompanied many ISDN rollouts--where services were later and pricier than anyone expected.
Analysts also contend that the widescale consumerization of DSL will force today's $1,000 price for a pair of rate-adaptive ADSL modems and accompanying central office equipment to drop to $500 to $650.
Finally, perhaps up to 40 percent of the 160 million copper lines deployed in the United States may need to be modified to accommodate DSL--at a cost some experts put at $100 per loop.
The Technological Heart Compounding these business and infrastructure challenges is the fact that two key modulation techniques are competing in the DSL space. The DMT (Discrete Multitone) technology standard, put forward by ANSI (American National Standards Institute) and the European Telecommunications Standards Institute, provides for feature-rich transmission up to 6.144 Mbps from central office to customer. Upstream transmission is set at 640 Kbps. A simpler and more readily available approach is the CAP (Carrierless Amplitude and Phase) modulation technique fostered by Paradyne, Lucent, Bell Atlantic and NYNEX.
A recent merger between competing vendors, Westell Technologies and Amati Communications Corp., is expected to tilt industry momentum toward DMT. Amati also recently announced a cross-licensing agreement to share patented DMT technology with Alcatel. Taylor estimates that CAP/DMT momentum is split 50-50 in the United States, with DMT enjoying an edge internationally.
Time to market is an important issue. While CAP is implemented in products, DMT products aren't expected to hit the market until year's end, with services based on those products emerging in the first half of next year.
The Applications Conundrum Important questions persist about which business applications are best-suited to the high performance promised by DSL. Some applications--browsers, for example--have had problems in the past with regard to taking advantage of increased bandwidth. However, work is under way in the IETF (Internet Engineering Task Force) to optimize TCP for high-speed and satellite transmission. Taylor cautions that both software and hardware need to be optimized to take advantage of DSL. He points out, for instance, that some PC communication ports don't support more than 1 Mbps.
What should businesses do until such issues are resolved? Ask questions: Find out what DSL provisioning is occurring in your service area or areas. Examine costs and payback, and budget considerations entailed in DSL. Consider turning over some of the technology risk factors--CAP versus DMT--to the carrier in the form of an outsourcing agreement. Rethink remote user strategies with DSL in mind. Make pointed inquiries about line quality in your service area. Draft desktop buying guidelines that ensure that remote access desktops purchased will support multimegabit speeds. Examine applications, and determine what technology provides the best access fit--be it DSL, T1, ISDN or yet another answer.
Saroja Girishankar is executive industry editor at InternetWeek. She can be reached at sgirisha@cmp.com. Christy Hudgins-Bonafield contributed to this article. She can be reached at cbonafield@nwc.com. |
