Server: Netscape-Enterprise/2.0a Date: Mon, 29 Dec 1997 20:57:05 GMT Accept-ranges: bytes Last-modified: Mon, 01 Dec 1997 20:23:04 GMT Content-length: 14150 Content-type: text/html @Home Network

Company Information

@Home Network Text Links Below

 

TOP
5 MYTHS
OF xDSL

xDSL vs. @HOME'S HYBRID-FIBER-COAXIAL (HFC) CABLE MODEM NETWORK: THE FACTS

The cable industry's hybrid-fiber-coaxial (HFC)/cable modem networks are proven commercial products that deliver high performance, reliable service, and great value. As of November 1997, approximately 100,000 customers in North America are actively using the Internet everyday at speeds never before possible, using HFC/cable modem services. As a result of its breakthrough integrated product offering and cable industry partnerships, @Home Network has over 40% share (and growing) of this explosively expanding market.

Recently a few companies have announced trials or limited commercial deployments of Digital Subscriber Line (DSL) technology. While many incompatible variants of DSL exist (e.g., SDSL, IDSL, CDSL, HDSL, ADSL, VDSL), the most commonly discussed are ADSL (Asymmetric DSL) and HDSL (High-bit rate DSL). In the excitement of limited deployments, a couple of myths have been promulgated in an attempt to forward xDSL's nascent cause - 1) xDSL can be deployed on existing copper wiring with no upgrades required, and 2) xDSL's "dedicated" connection enables better performance than cable modems.

In addition, while xDSL technology is a potential advancement over current telecom capabilities like ISDN, it's worth noting that xDSL has fundamental technology limitations relative to HFC/cable modem systems. The balance of this note provides some facts to help objectively compare xDSL and HFC/cable modem offerings.

Plant upgrades for xDSL and cable modems: Both require some work
While proponents of xDSL tout its use of copper telephone wiring, it is a mistake to assume that it's a "free lunch" and that substantive plant upgrades of the telecom infrastructure will not be required. To start, carriers must remove load coils, bridge taps and splits, all of which are scattered across their plant today. Load coils block ADSL's high frequencies, and bridge taps reduce the signal and produce reflections on the line. Telecom carriers will also likely have to rehab outside plant (OSP) components such as splice modules, cross connects and terminals, in an attempt to minimize near- and far-end crosstalk and to correct resistive and capacitive faults. America's Network (a magazine of public network providers) recently confirmed "only a small portion of the OSP facilities across the U.S. can dependably support ADSL or any xDSL technology ..." Consistent with this finding, one of the RBOCs recently shelved plans for a 5,000 subscriber ADSL deployment when it found only about 50% of its copper phone loops capable of handling ADSL. Old and/or abnormal premise wiring also presents a problem. Given that much in-home wiring has been in place for more than 30 years and has been deregulated for more than 15 years, "it appears that internal wiring in many homes will have to be replaced in order to deliver ADSL" (LANTimes). Finally, the networking infrastructure behind xDSL must be built from scratch, which requires capital expenditure for new equipment for each incompatible xDSL variant used in each telephone central office (CO) where it is deployed.

The cable infrastructure also requires upgrades to enable service. Infrastructure that is already HFC has to be "2-way activated". Upgrade of non-HFC infrastructure has been a focus of aggressive cable company deployments for the last several years as part of ongoing initiatives to improve quality, expand channel offerings, and enable other digital services. Currently over 10 million homes in North America are 2-way HFC.

"Dedicated" vs. shared access: Everything is shared at some level -- It's what's behind it that counts...
Some xDSL advocates have attempted to contrast HFC/cable modem's shared high-speed neighborhood coax segment with xDSL's lower speed "dedicated" connection. Accompanying the contrast is often the allegation that xDSL will have superior performance when more customers are added to cable systems because "the more [cable modem] users online, the lower the speeds."

Unfortunately this simplistic view overlooks a fundamental premise of modern data networks - everything is shared at some level. ADSL in most viable implementations, for example, becomes a shared connection immediately after the local loop is terminated at the telephone CO when many lines interconnect to the so called DSLAM (Digital Subscriber Line Access Multiplexer). After concentration at the DSLAM, shared data connections are passed through one or more ISP interconnects and ultimately to the Internet backbone. Typical concentration rates discussed for xDSL implementations are 100:1 (meaning that 100 users with "1.5 Mbps-rated" ADSL service would share a single 1.5 Mbps T-1 ISP connection; further concentration may occur before eventually reaching an Internet backbone).

Given that xDSL and cable modems both integrate into shared data networks, the real performance drivers are: 1) the ability to scale with the number of customers, 2) "effective bandwidth" provided per customer, and 3) backbone connectivity.

HFC/cable modem systems proven scalability. The shared high-speed cable modem architecture is similar to local area networks (LANs) used in offices everywhere. The approach is particularly efficient and well-suited to the "bursty" nature of Internet traffic, where surfing tends to consume bandwidth in discrete bursts intermixed with periods of inactivity. With a high-speed connection and Internet traffic, it is not correct to estimate bandwidth by simply dividing by the number of users. For example, a recent CableLabs test of cable modem systems demonstrated no detectable performance impact for up to 200 concurrent web-surfing users on a single coax segment.

Given that cable system "nodes" (the neighborhood-level junctions where signals convert from fiber optic to coax) typically cover 500-1,500 homes and not all customers are active at the same time - there's ample room for substantial penetration growth without any performance concerns. When needed, however, HFC's "fiber-to-the-neighborhood" approach has a great deal of flexibility for scaling local bandwidth with even very high penetration levels. @Home and its cable partners can pragmatically and economically increase local capacity many times over in several ways, including segmenting cable nodes and allocating additional frequency for @Home service (e.g., another 6 Mhz channel out of the typical 80-120 available in modern systems). These approaches allow fine-tuning the amount of bandwidth available, as well as the number of users accessing that bandwidth.

More "effective bandwidth" per customer. Each @Home customer benefits from much higher "effective bandwidth" than any competitive offering or technology available. The high effective bandwidth is the result of combining the cable industry's extensive fiber-based networks, @Home's high-speed national backbone, and advanced technologies that @Home has developed and integrated to enable sustained end-to-end high performance. These core @Home technologies - caching, replication, and multicast - increase effective bandwidth by improving efficiency and reducing redundant data traffic.

Superior backbone connectivity. For top performance and reliability, @Home has deployed its own high-speed national backbone (effectively a "parallel Internet"). In addition, @Home is a "Tier 1" service provider directly peering with other top ISPs at key Internet network access points. The combination ensures @Home customers the highest performance and least congestion possible.

No known xDSL providers have made the commitment to sustained end-to-end performance that @Home has. If xDSL users just hook up to an ISP, they'll likely run into the same overloaded Internet traffic jams as others experience today.

Limitations of xDSL technology
In addition to the substantially lower bandwidth-carrying capabilities of copper wiring relative to coax (e.g., 1.5 Mbps maximum for copper at typical distances vs. 30 Mbps from just one 6Mhz channel out of 550-750 Mhz total available), xDSL has some marked technical limitations. The most notable limitation is severe distance attenuation. According to the Chairman of the ADSL Forum, "with ADSL alone only 50% of homes and businesses will be able to be reached because they are too far from their respective COs." Distance attenuation rears its head when ADSL upstream bandwidth drops precipitously from its 640Kbps maximum down to just 64 Kbps (or less) for any customers beyond 12,000 feet from the telephone CO.

Another exacerbating problem is near-end cross talk (NEXT). NEXT is noise produced by T-1 lines and other copper lines using DSL or ISDN. Considering that the copper wires connecting a neighborhood to the main office are often bundled together when buried and only separate once the bundle reaches the neighborhood, the impact can be substantial. If ADSL were widely used, a large number of lines all with vexing NEXT noise would converge in immediate vicinity near the main office; crosstalk severely limits the ability to provide a consumer-priced, high penetration offering.

Finally, xDSL deployments have suffered from technology immaturity. For example, today 40 ADSL equipment makers sell six versions of ADSL in two incompatible modulation "standards." xDSL modems are typically much higher priced than cable modems due to low volume and lack of standards (e.g., $500 - $2,000 each). In contrast, cable modems have standardized on the interoperable MCNS specification. The MCNS cable modem standard paves the way for fast rollout, affordable equipment, and plug-and-play consumer devices.




In summary, proven HFC/cable modem technology has key advantages over xDSL. RBOCs don't have a "free lunch" on plant upgrades. The cable industry has a sizable lead on commercial offerings, and is aggressively moving forward on remaining HFC/2-way deployments. xDSL's "dedicated" connection becomes a shared connection quickly, like every data network. The true performance drivers are ability to scale with the number of customers, "effective bandwidth," and backbone connectivity. @Home was designed from the ground up to meet these challenges, and HFC's inherent advantages combined with @Home's advanced technologies make high speed data services a reality today. Both xDSL and cable modem service will bring advances to customers, but cable's HFC advantages with @Home's integrated solution ensure it will dominate.

Top of Page


Company | @Home | @Work | Support | Advertising | Store | Sign Up | Home Page