Cable Versus Dsl

Home , Cable modem, Digital subscriber line

53-10-60

DATA COMMUNICATIONS MANAGEMENT CABLE VERSUS DSL

John R. Vacca

INSIDE DSL; Cable Modems; ADSL; CDSL; G.Lite; HDSL; IDSL; RADSL; SDSL; VDSL; POTS; DSL and Cable Modem Rollouts; High-Speed Data Entry; Buying DSL Service; Installing DSL; Security Problems, Residential Users, Telecommuters, DSL System Components; DSL Network; DSL Hubs

INTRODUCTION Internet access via cable modem has become available in many residential areas over the past few years. Cable has the capacity to transmit data at speeds as fast as Digital Subscriber Line (DSL) when conﬁgured prop- erly and under optimal conditions. Due to the fact that cable lines are not available in the vast majority of commercial districts, cable does not compete with DSL in the enterprise market at all, in most cases. Cable was designed for residential use, and in some cases may be a cost-effective solution for residential high-bandwidth Internet access. Therefore, the challenge of cable versus DSL is primarily in the residential and telecommuter markets. With that in mind, and before continuing with the theme of this article (cable vs. DSL), one can take a look at the technology issues ﬁrst, and then some basic terminology.

TECHNOLOGY ISSUES What is DSL? How does it work? What are the types of DSL? These are some of the questions this article will surely answer; as well as some of the pros and cons of the use of cable modems versus DSL. PAYOFF IDEA The article discusses the current state of cable DSL: What Is It? modem access versus DSL. It also examines how In essence, by using the existing tele- prevalent cable modem and DSL services are in major U.S. markets. A comparison of the two phone cabling infrastructure, DSL is technologies with regard to speed, cost, etc., are a technology backed by telephone presented. The article also covers the planning enterprises that provides high-band- implications or considerations for the enterprise width services to the home and en- network manager (e.g., to support telecommut- terprise. Because DSL utilizes a ing employees, etc.). Finally, the article discusses future directions for cable modems and DSL.

DATA COMMUNICATIONS MANAGEMENT greater range of frequencies than ordinary dial-up services (allowing for a super-fast connection), this high bandwidth is possible. For most providers, this technology is still in the early stages of rollout.

How Does It Work? The general idea behind DSL technology is relatively easy to grasp, although it is rather sophisticated. As previously mentioned, DSL utilizes a large range of frequencies, which means a higher bandwidth and a fast connection. For example, consider this: only a small fraction of one’s telephone line capacity (bandwidth) is being used (that being only the low frequencies) when making an ordinary telephone call. By transport- ing data in the higher frequencies, DSL takes advantage of this idle bandwidth. This results in making it possible to talk on the phone and be on the Internet simultaneously.

DSL Types As shown in Exhibit 1, there are several competing forms of DSL, each adapted to speciﬁc needs in the marketplace. Some forms of DSL are widely used standards, some are proprietary, and some are simply theo- retical models. They can best be categorized within the modulation methods used to encode data. Exhibit 1 shows different types of DSL technologies. These technologies are sometimes collectively referred to as xDSL. As explained in Exhibit 1, Asymmetric Digital Subscriber Line (or ADSL) is the most popular form of DSL technology. The fact that the upstream and downstream bandwidth is asymmetric, or uneven, is the key to ADSL. In practice, the higher-speed path will be the bandwidth from the ISP to the user (downstream). This is mainly due to the desire to ac- commodate the typical Internet usage pattern, where the majority of data is being sent to the user (Web pages, graphics, programs, and video) with minimal upload capacity required (keystrokes and mouse clicks). Speeds typically range from 144 Kbps to 1.1 Mbps downstream, and range from 144 Kbps to 1.5 Mbps upstream. Exhibit 1 also shows that there are other forms of DSL as well: ADSL Lite; Consumer Digital Subscriber Line (CDSL is a proprietary technology trademarked by Rockwell International), G.Lite, HDSL, IDSL, RADSL, SD- SL, and VDSL. Many of these forms are just starting to become available through the telephone enterprises, and some of them have just complet- ed the testing/development stages.

DSL versus Cable Modems Cable modems, although capable of high potential access speeds, also have drawbacks. Primarily, the signal is shared between the subscribers

CABLE VERSUS DSL

EXHIBIT 1 — The Many Flavors of DSL

Type Description

ADSL Asymmetric digital subscriber line. The most common standard, it theoretically offers 1.5 Mbps to 8 Mbps downstream and 16 Mbps to 640 Kbps upstream speeds. HDSL, HDSL2 These two high data rate DSLs are symmetric services capable of 1.5 Mbps and 2.048 Mbps speeds, respectively. HDSL requires two or three wire pairs; HDSL2, just one pair. IDSL ISDN DSL. Similar to ISDN, it allows you to use existing ISDN equipment. But the maximum speed in both directions is 144 Kbps. RADSL An asymmetric service, rate adaptive DSL promises to provide between 600 Kbps to 8 Mbps downstream and 128 Kbps to 1 Mbps upstream speeds, while offering simultaneous voice service. RADSL can dynamically adjust to line conditions. SDSL Symmetric DSL, a popular alternative to ADSL, is offered by NorthPoint and various ISPs. The service promises two-way 768 Kbps access. G.Lite G.Lite is user installable and provides speeds of 1.544 Mbps downstream (UADSL/ and 512K bps upstream. Backed by many hardware vendors and the DSL-Lite) Universal ADSL Working Group. VDSL The fastest and newest DSL on the block, very high speed DSL is an asymmetric service, offering speeds of 12.9 Mbps to 52.8 Mbps. in a speciﬁc area, and the technology is broadcast oriented. Less bandwidth is available to each subscriber as the number of subscribers in- creases in that area. Moreover, cable access offers the user no choice regarding providers, and is only available in a limited area. It is available only through the cable enterprise, which has little experience with Inter- net services.

CURRENT STATE OF CABLE MODEM ACCESS VERSUS DSL So what is holding up DSL and cable modem access? If one has been told that the lack of widely accepted standards is the main factor delaying cable modem and xDSL access, think again. Lack of standards is, at most, a minor factor in the delay of widespread access. Because solid standards have not been established, xDSL and cable modem access are being held up. So, one needs to put things in perspective although it is certainly true that the lack of standards is having some effect. Cable modem and xDSL services are dedicated subscriber services. One uses a cable modem or xDSL box to connect to one place — the service provider. One does not carry it around like a modem. Where you expect to move the device from place to place or to a different service provider (as with a 56 K or 33.6 K modem), standards only affect the end user in cases. A cable modem or xDSL box is like a television cable box (which is never moved — one buys or leases it with the service and it stays with the service) — not like a modem. By allowing vendor competition, it is true that standards will lower hardware costs a bit. This, in turn, will lower equipment costs, but not by much. The important thing

DATA COMMUNICATIONS MANAGEMENT is that if the provider can offer service at a price one likes, that is what matters most. The fact that the provider will be able to buy the box, now being leased for $10 a month, for $100 less next year is simply not that much of an issue. Anyway, it is probably a small fraction of the overall cost of providing the service. The history of cable modem access by Cablevision on western Long Island over the last several years tends to support this. The original beta access used a speciﬁc brand of cable modem (which was provided with the service). When the beta program ended, the original equipment was discarded in favor of a newer, higher performance product made by Lan- City. This did not affect end users much — they just returned the original modems and installed the new ones. The current service provides a 10 Mbps Ethernet connection for under $70 per month (this includes the lease of the cable modem unit, connection, and ISP service) for residential customers. This is three to ﬁve times faster than a full T1. The LanCity cable modem follows the only real standard that matters so far — it con- nects to the user’s PC using standard 10 Mbps Ethernet, like most other xDSL units and cable modems. One has to look elsewhere than the lack of widely accepted standards if one wants to know why xDSL and cable modem service is slow in becoming pervasive. Maybe one should wonder if the cable enterprises are simply afraid of moving into a new technology they do not understand; and if the telcos are not that eager to access a DSL service in direct competition to their data T1 and T3 services (and with a lower price tag); and whether anybody has enough backbone capacity to support the large numbers of high-speed customers the new service would attract. Inci- dentally, Cablevision claims no immediate plans to offer enterprise services. They are not interested in going outside their traditional market to offer service to more demanding enterprise customers (who generally use a higher percentage of their available bandwidth for a larger part of each day than residential customers). Nevertheless, the battle for broadband data services to the home and enterprise has been played out as a battle between the telephone enterprise (telcos) and the cable operators. The battle has been fueled by “arms merchants,” the silicon suppliers and xDSL and cable modem box vendors, selling both hype and product to both sides. With over 360,000 in commercial operation in North America alone and the entire industry rallying around CableLabs-promulgated standards and retail distribution models, cable modems have taken the early lead. A tremendous lure for the cable operators is the potential for remote LAN access and IP-telephony on top of Internet access. Meanwhile, in an attempt to solidify the industry and catch up with cable deployments, the DSL community is moving toward a low-cost, splitterless G.Lite standard. G.Lite could be an easily deployable, low-cost technology that carries a dedicated IP pipe- line on top of the twisted pair voice connection leading into virtually ev-

CABLE VERSUS DSL ery home. Or, G.Lite could be just another vegetable in DSL alphabet soup. The major regional Bell operating companies (RBOCs) have an- nounced DSL access plans for into the year 2000, despite a myriad of technical and enterprise hurdles. So, what is real and will there be a winner in this battle for broadband in the local loop? Well, a recent bitter struggle among equipment manufacturers is threatening new industry rules that could make broadband in the local loop and high-speed Internet connections less expensive, easier to use, and available to far more people.

DSL Progress Is Thwarted The infighting centers on the so-called G.Lite standard, which is designed to simplify cable modems used for digital subscriber line (DSL)1 connections. Thus, manufacturers need to make products that can work together seamlessly to speed wide-scale adoption of DSL, because the market is littered with different technologies. This type of conflict is common with emerging technologies, but rare- ly are the stakes as high as they are today. High-speed access is largely associated with the growth of the Internet itself, and billions of potential dollars hang in the balance for those enterprises that come up with the winning solutions to break the widespread bottlenecks across the Web. Moreover, the rhetoric is particularly rancorous in the G.Lite debate. Critics say G.Lite modems do not work well with current technology, according to service providers. And some manufacturers are now fighting to see their own products adopted as the industry standard. In some cases, these technologies cannot even communicate. It is not just that one needs a G.Lite modem. One needs a G.Lite modem from a specific manufacturer. And that is the same as no standard at all. While equipment makers argue over how to best implement the G.Lite standard, competition from high-speed cable modems is intensifying. The cable industry has already established a standard set of rules for cable modems, giving the group a leg up on its DSL rivals. According to TeleChoice, about 270,000 DSL lines were in use by the end of the second quarter of 1999, with 84 percent of these in residences. Cable modems have reached well over 2 million users. Just as telephone enterprises and Internet service providers like Amer- ica Online and Prodigy ramp up their marketing machines to push high- speed Net services across the country, the lack of a universally accepted standard could slow the spread of DSL. To have an absolute standard that leaves no variation in the way that equipment vendors can make their equipment is not practical.

All In One. So that different enterprises create products that can work with each other, many industries push for technology standards, or rules.

DATA COMMUNICATIONS MANAGEMENT

For example, VHS is a standard for videotape. Regardless of the enterprise that manufactured the VCR, a VHS tape cassette will play on any VCR machine that is built to handle that certain type of tape. Standards are critical in the communications world. If a consumer buys a modem, it must work with his or her computer, as well as with the ISP. One cannot get to the point of popularity reached by ordinary analog modems without a standard. For example, people should be able to buy a computer at CompUSA, etc., take it home and plug it in, and have it work with their ISP. Several years ago, the Internet and computer industries coalesced around the G.Lite standard for consumer DSL. It was ratiﬁed2 by the In- ternational Telecommunications Union as the worldwide set of ofﬁcial technical guidelines. Rather than wait for a phone enterprise technician to install a modem, the standard is aimed at making DSL relatively cheap, as it would allow a consumer to buy a modem and plug it into a PC. The challenge is to make sure that modems and other equipment made under the G.Lite standard all work together.

Lab Tests. The University of New Hampshire has set up a lab where enterprises like Intel, Alcatel, Lucent, and others can test their modems against each other and against other products — all under the auspices of the ADSL Forum. Nevertheless, some enterprises already say they comply with the standard, and some computer makers (led by Compaq and Dell) have already begun shipping machines with standardized DSL modems that conform to G.Lite rules. However, the modems still will not work with all ISPs and telephone enterprise equipment. The big telephone enterprises (the enterprises that will invest the most money in DSL equipment), for their part, are testing G.Lite technology and hope that manufacturers will end their differences as soon as possible. Most phone carriers are using different versions of DSL on an interim basis in the meantime. This has taken the interoperability process to a whole new level. It will take time to sort it out. So, how prevalent are cable modems and DSL in major U.S. markets?

PREVALENCE OF CABLE MODEMS AND DSL IN MAJOR UNITED STATES MARKETS Cable modems and DSL are somewhat prevalent in major U.S. markets. They are, nevertheless, a quantum leap in Internet access, forever putting an end to the world wide wait. All the hype about instant information on the Internet does not mean much when one has a Web page downloading so slowly that one has to keep busy by ﬂipping through the Sunday paper or glancing at the head-

CABLE VERSUS DSL lines on CNN. There is an apt, if overused, name for this trickling of info: the world wide wait. As previously discussed, now two competing technologies (cable modems and DSL) would like to bring high-speed Inter- net access into the home, ending the wait with speeds up to 200 times faster than what is possible with standard modems. For as little as $30 per month, these next-generation methods of connecting to the Internet promise to transform the cyberspace experience. Think video on demand or live stock tickers. Both are possible, if one has the speed. Chances are, however, such access is not available yet; cable enterprises, phone enterprises, and Internet service providers are in the early stages of offering cable modems and DSL. One is lucky to have access; if there is a choice between the two, one is very, very lucky. DSL launch- es are under way or planned by Bell Atlantic, GTE, and US West, and with cable enterprises expecting competition from DSL, one is likely to hear about high-speed access coming to the home of a friend — and maybe even one’s own home. Skeptical? Witness AT&T’s acquisition of cable-TV giant Tele-Commu- nications, Inc., a merger spurred by the prospect of offering new services (and packaging existing ones) through high-speed lines into the home. The deal was a $59 billion vote of conﬁdence in the Internet’s transfor- mation from a frequently slow and frustrating means of communication into a mature, consumer-friendly medium. Other providers now have a newly energized mega-competitor in this arena, and that is likely to mean a faster rollout of high-speed access. But just in case one thinks this is another overhyped upgrade, like the much-ballyhooed 56-kilobits-per-second (Kbps) modem, think again. The advances in standard modems have been incremental: from 14.4 kbps to 28.8 Kbps to 33.6 Kbps to the current champ, 56 Kbps. Super- charge all that, and imagine the switch from 33.6 Kbps to, say, 1.5 mega- bits per second. That is 45 times faster, meaning a 5.6-megabyte game would take about 30 seconds to download, instead of 22 minutes. Or the hundreds of pages one is viewing in a single session of Web browsing will appear without delay, as if the pages were stored on one’s hard drive. Aside from speed, both cable modems and DSL offer another advantage: an always on connection to the Internet that does not tie up the phone line.3 There is a lot of pent-up demand for this. Most people will take the ﬁrst thing that comes along. To check on availability, call a local cable enterprise or (for DSL) the phone enterprises and Internet service providers in the area. Costs can vary wildly, since most high-speed-access providers currently face no competition. Cable modem access typically runs about $30 per month; DSL is costlier, and its pricing is more complex.4 For cable modem access, local cable systems package an Internet service provider (@Home is the leader) with the speedy line into the home. With DSL, one is likely

DATA COMMUNICATIONS MANAGEMENT to have the option of choosing an Internet service provider and the level of speed, from 256 Kbps on up. Few ISPs offer DSL or cable modem op- tions yet, so check with the current provider to see if high-speed access is available. Consider the DSL offerings of US West, now available in 50 cities, including Denver, Phoenix, Salt Lake City, and Minneapolis–Saint Paul. For $30 per month, one gets DSL that runs at 256 Kbps, or one can combine it with US West Internet access for $48.84. Want more speed? Like 512 Kbps? That will cost $54 a month, plus $28.84 for Internet access. Other DSL providers have similar plans, commonly offered in tiers for casual users and enterprise customers. For the time being, at least, cable modems are the better value. Expect to pay installation fees of up to $140 for either one; a technician must visit the home to handle the wiring. A high-speed access modem can cost as much as $200, but some providers may have a rental option or even include modem fees in the monthly rate. Problems? The biggest headache is that it may not yet be available in one’s area. Nevertheless, will DSL service for faster access to the Internet catch on with enterprise users? Analysts say it will, but users are not as sure.

Unsure Users Many telecom managers are still waiting for the success stories with DSL technology. Several respected consulting enterprises have projected phe- nomenal growth in DSL technology and service. The Gartner Group, Inc., in Stamford, Conneticut, predicts annual growth rates of more than 400 percent, with more than 2 million DSL lines installed by 2002 — up from fewer than 60,000 now. Analysts’ conﬁdence in the market’s growth is based on the strong need for faster connections to the Internet sought by consumers and enterprise users who telecommute or work in remote locations.

Beneﬁts The most-discussed variant of DSL, Asymmetric DSL (ADSL), boasts downstream speeds that are more than 40 times faster than 56 Kbps. connections. But analysts say the upstream speeds are much slower, making ADSL less attractive for enterprise users who need to push large ﬁles to colleagues and enterprise partners. Several carriers are deploying or testing Symmetric DSL (SDSL) to solve the need for fast upstream speeds using equipment by start-up vendors such as AccessLan Communications, Inc., in San Jose, California; or Copper Mountain Networks, Inc., in Palo Alto, California. And established networking vendors such as Cisco Systems, Inc., and Bay Net- works, Inc., have entered the arena.

CABLE VERSUS DSL

EXHIBIT 2 — DSL versus Cable

Beneﬁt DSL Cable Modem

Consistent download speed Yes No Consistent upload speed Yes No Immune to degradation from neighborhood usage Yes No Secure, exclusively dedicated line with isolated data Yes No Carrier guarantees performance Yes No

Competitors in the SDSL market hope to attract customers with lower price and convenience of installation, compared with installing T1 lines. AccessLan will give enterprises a 1.5 Mbps SDSL connection at lower cost than a typical T1 (1.544 Mbps) connection from a telephone enterprise. Therefore, if DSL spreads as quickly as everyone thinks it will, there will be pressure on telcos to eat their T1 prices. Another factor is how long it takes an enterprise to install its own on- premise SDSL equipment, compared with waiting for a T1 connection. With SDSL, a small router is installed at the enterprise and an access con- centrator is installed in a carrier’s central ofﬁce. The two are connected by existing twisted-pair phone cable. The process may take only a day, compared with waiting weeks for a carrier to install a T1 line. Most users would consider using cable modems for faster speeds, if they were available. Cable modems will not have as much growth among enterprises as DSL, partly because ofﬁce parks often are not wired for cable.

COMPARISON OF CABLE MODEM VERSUS DSL: SPEED AND COST When compared to DSL, the cable modem has few advantages and several disadvantages for Internet access. DSL, on the other hand, has many advantages. Unlike ISDN, DSL is not metered by the minute. Better yet, DSL is truly an always on service, which makes it suitable for a wider range of applications, including interconnecting LANs, small Web site hosting, videoconferencing, and connecting branch ofﬁces (see Exhibits 2 and 3).

Performance In every cable modem network neighborhood, hundreds (perhaps thou- sands) of households all share a common transmission medium. Down- load speed could be high if conditions are optimal with absolutely no activity from neighbors. However, typical download speeds in many neighborhoods are no higher than 400 Kbps range during real-world usage. Upload speeds are even slower because the cable modem network

DATA COMMUNICATIONS MANAGEMENT

EXHIBIT 3 — High-Speed Services Compared

ISDN T1 DSL

Speed 128 Kbps 1.544 Mbps 128 Kbps to 1.544 Mbps Cost per month $50–$110 for $700–$2,400 $78–$314, (including moderate usage depending on Internet service) speed Notes Per-minute charges Price can vary with Wide spectrum of make full-time mileage, usage, services; distance connections and service level limitations; impractical incompatible versions is optimized in one direction only and, depending on cable provider, often capped.

Value As more users pile onto cable modem services, performance goes down rapidly, diminishing the value of the service. In response, many cable modem providers around the country now cap subscriber outbound speeds at 128 Kbps (equivalent to about two 56 Kbps modems). DSL is immune to both real and artiﬁcial speed restrictions. The quality and value of DSL remains constant for the lifetime of the service, while cable modem Internet quality degrades and loses value over time. Combine that with a yearly term imposed by the cable modem enterprise and one may be locked into a losing investment.

Security Aside from performance variations and diminished quality, the shared cable modem network also introduces concern for security. Data from one household passes over the same wire connecting others, potentially exposing credit card, e-mail, and other sensitive data. For this reason, users either ﬁnd cable Internet access unacceptable, or must arrange en- crypted software tunneling to secure networks, which impedes performance.

No Surprise Here Not surprisingly, DSL use is expected to boom. According to market research analysts at TeleChoice, the number of installed DSL lines will grow from 433,000 in 1999 to 1.015 million in 2001 to 3.46 million by 2003. DSL will do well among people working at home (telecommuters) and in urban areas where cable services are not widely deployed.

Residential Users. Residential cable modem service can be a good value for high-speed access. Still, there are some limitations for residential

CABLE VERSUS DSL cable modem service. Cable is a shared medium — unlike DSL, which offers a dedicated line for each user. When too many users in one neighborhood try to share the same cable, performance can suffer dramatically as users compete for the limited resources that one cable can afford. For this reason, residential cable customers are forbidden to telecommute, host Web sites, or use video teleconferencing or any other bandwidth-intensive applications. As previously stated, using a shared medium also creates security problems; and, cable modem users are much more vulnerable to data interception, unauthorized monitoring, and hacking from other users along the same cable network. While there is a market for residential cable modem service, obviously there can be signiﬁcant drawbacks.

Telecommuters. Telecommuters using cable modem service are required to pay a rate designed for the few enterprises that have cable modem service available. Service packages start at almost double what a DSL telecommuter would pay for the same speeds. Additionally, cable telecommuters are subject to the same security and bandwidth problems that residential cable customers face. Therefore, cable modem service is not a particularly attractive option for these customers.

Choices Although the number of enterprises supplying DSL service and Internet access is increasing, would-be users face a confusing range of choices about availability, price, equipment, and conﬁguration. DSL service is hardly ubiquitous. Even in areas where DSL has been deployed, if one is too far from the phone enterprise’s central ofﬁce (where the switching hardware is located), one cannot get it or does not get the maximum possible speed. Prices, while reasonable for enterprises, are steep for in- dividuals. 128 Kbps service can cost as little as $78 per month, while 1 Mbps to 1.544 Mbps service can go for several hundred dollars. Equip- ment is not widely standardized, and there are a number of DSL variants to contend with (see Exhibit 1). So, what are the planning implications for the enterprise network manager?

PLANNING IMPLICATIONS FOR THE ENTERPRISE NETWORK MANAGER In the enterprise network manager, operator, and service provider world, it is now an accepted, inescapable conclusion that xDSL technologies will one day be installed on a global basis, in vast quantities. Despite other technology developments (not excluding recent announcements about the commercial viability of sending data down power lines), it is still the only viable technology capable of substantially increasing bandwidth on the local access loops without a substantial overhaul. These

DATA COMMUNICATIONS MANAGEMENT copper loops are ubiquitous over every home and it is unlikely that any mass-scale upgrade to ﬁber will happen within the next 13 years. The primary driver for xDSL is high-speed Internet service deployment to residential customers. So far, there have been two small commercial deployments and no less than 66 trials around the world, trying to prove that Asynchronous DSL (ADSL) can provide a downstream connection of up to 8 Mbps and upstream connection of up to 1 Mbps over the existing telephony copper pair. Both IP and ATM network architectures are currently on trial. ATM is considered the probable choice in the future. Initial trials have used stand-alone ADSL modems with discrete IP interfaces for reasons of time- to-market and availability, and these are evolving to highly integrated digital subscriber loop access multiplexer (DSLAM) solutions. The complete network typically consists of a core that is based around SDH (Synchronous Digital Hierarchy), wrapped with broadband ATM switches, ATM access switches, and DSLAMs. The DSLAMs provide the individual xDSL lines out to the customer premises and integrate with the existing POTS network connections. But even after having been in trial for more than 18 months, few operators seem close to commercial rollouts. What is taking them so long? Some of the main barriers to large-scale adoption include the continuing standards battle and the lack of interoperability. The existence of two de facto standards, namely DMT and CAP, provides network operators with a dilemma: deciding which to adopt before a leader is clearly iden- tiﬁable in case the choice ends up as the Betamax of the standards. There are two standards camps, one apparently led by Israel’s Amati, which has developed modems using discrete multitone technology (DMT) for the line coding. Motorola and Alcatel are among the major manufacturers who have developed DMT modems. The second camp has adopted a technology developed by the former AT&T Paradyne, which champi- oned carrierless amplitude modulation/phase modulation (CAP) for the line coding. Westell,5 the enterprise that made the ADSL modems being used in Bell Atlantic’s Virginia trial, uses CAP. With no clear stance being taken, a general wait-and-see attitude is developing. Lack of interoperability is also becoming a matter of concern. The current xDSL units on the market, from a variety of vendors, supporting different standards, do not interoperate. Interoperability between vendors’ equipment is key to mass rollout and provisioning. The ultimate goal is for end customers to be able to purchase their own ADSL termination units, much the same way as people buy off-the-shelf analog modems today. For this to happen, the ADSL termination unit technology needs to mature to the point where it is as simple as a modem to install and operate. Operators are still deciding on suitable end-to-end architectures. Once these infrastructures have been agreed upon and the business and technology are in place, services can then be rolled out in volume. These new

CABLE VERSUS DSL architectures need to address the integration of network, service, and enterprise management of these new broadband services. Without these seamless, end-to-end management systems in place, it is difﬁcult to see how operators will be able to make any proﬁt on broadband services.

Enterprise Network Management Capabilities The new broadband services being deployed by enterprise network managers, operators, and service providers are typically delivered over very complex network environments, which include legacy equipment with primitive management capabilities and newer systems incorporating sophisticated telecommunications management network (TMN)-based element management. New network management infrastructures are required to mold these disparate information sources into a cohesive view of the end-to-end services being delivered. This can be complicated by the large scale of such services — typically involving hundreds of thou- sands or millions of network elements. The major requirements can be summarized as follows:

• Cost-effective, performance scalability: a system with the ability to manage small pilot networks and expand as the network grows without performance degradation. • Multiple protocol support: standards-based and proprietary management protocols must be supported in a transparent manner to the network applications. • Powerful event management: as the network grows, the number of events grows exponentially. The ability to manage this volume of events in a way that helps the operator make sense of the status of the network is mandatory. • Seamless application interworking: operators will use a number of different applications to manage aspects of the network, but they require the ability to move easily between applications, to investigate problems or conﬁgure service for a particular customer, for example. • User access management: the information within the management system must be protected, but it must also be feasible to partition the data, and the access to it, in a way that supports operational processes.

Enterprise Network Management Architecture Some of the key management challenges when installing xDSL networks include potential element volumes. xDSL network elements are extremely complex devices that require realtime management. As element volumes increase exponentially, operators are faced with the challenges of implementing end-to-end systems that can manage millions of highly complex access elements.

DATA COMMUNICATIONS MANAGEMENT

As an example, the 1994 BT video-on-demand (VOD) trial supported 2000 ADSL lines, equating to 4000 access elements. Although this was classified as a small marketing trial, it is still the largest single xDSL installation in the world. This network generated a tremendous amount of management traffic, primarily with SNMP and performance monitoring information, all of which had to be processed in realtime by the central management workstation. Additionally, xDSL equipment, by virtue of its inherent built-in intelli- gence, generates a tremendous amount of management traffic. This information — be it alarms, performance monitoring information, configuration data, diagnostic commands, or inventory data — must be processed, stored, prioritized, formatted, and displayed by the central management console. A single-element management workstation cannot scale to support any major network growth. Element and network complexity is also an issue that operators need to prepare for. xDSL equipment is becoming more complex with each generation. The latest generation uses rate-adaptive transmission, which causes a myriad of possible upstream and downstream speed permutations. Remote equipment now incorporates xDSL transmission hardware, IP routing equipment, and IP or ATM CPE interfaces into a single unit. In addition, it must be seamlessly integrated with the existing POTS network, leading to a wealth of configuration permutations, performance monitoring information, alarm and event reports, and diagnostic func- tions and inventory data being available to the network operator. A typical end-to-end network capable of delivering high-speed Inter- net, on-demand services, and data and voice from multiple service providers is extremely complex. Many issues can only be resolved with the use of integrated management applications and systems. This often results in a number of noncompatible element management systems being installed to manage different parts of the network, then integrated together to provide an overall end-to-end service management system. The even- tual goal is to provide a zero touch system to allow end customers to con- trol and manage a large part of their own service offering in realtime. xDSL equipment is normally installed over existing telephony circuits but the devices at each end of the connection contain POTS splitters that allow both the telephony circuit and the xDSL circuit to use a single copper pair. This requires the integration of a new xDSL management system with existing POTS systems. A range of issues also cloud the most efficient interface for the customer premises equipment. The two protocols in question are IP and ATM. Network operators are installing complex networking equipment into residential environments for the first time. This can lead to a multi- plicity of new problems. These problems range from units being tam- pered with and cables being disconnected, to PCs being reconfigured.

CABLE VERSUS DSL

This could result in a huge increase in the number of service-related calls and queries directed at the network operator. The operator has to be able to diagnose, isolate, and analyze problems from the central management system. Only when this type of network, system management, and con- trol is achieved will broadband services become commercially viable.

Viability The density of the technology must increase dramatically to enable operators to fit all the required ADSL termination units into their existing ex- change real estate. xDSL technologies are still costly, and this is holding back general acceptance due to the costs of equipment investment and the ongoing line rental charge. Prices in the range of $260 to $310 would start to approach an acceptable level for the benefits of this new broadband technology. But in most European countries, high-capacity broadband ATM backbone networks have yet to be deployed. Without this backbone infrastructure, there is little benefit to be gained from introducing an upgraded access network. In the infrastructure context, the questions of how some major European operators will play off ADSL against their substantial ISDN services will be an interesting debate. As it now stands, current versions of ADSL cannot co-exist with ISDN BRI lines. New versions of ADSL are being developed that will co-exist with ISDN, but as yet are only available in small trial volumes from one or two vendors. Telecommunication enterprises such as Deutsche Telekom, which bets its future on ISDN, are now involved in desperately trying to increase ISDN usage before ADSL establishes itself. The situation it faces (in common with many other operators in Europe) is that too much em- phasis on keenly priced ISDN will undermine ADSL; but, at the same time, not investing an adequate amount of resources in ADSL now might undermine its future as a broadband services provider. So, what does the future hold for DSL and cable modems?

FUTURE OF CABLE MODEMS AND DSL From the point of view of a commercial (proﬁtable — both for provider and user) service provider, xDSL can be considered an immature technology. xDSL has been proven an effective, reliable solution in many international trials, and is developing and progressing continually. This adds to the xDSL management complexities requiring the facilities for operators to download new algorithms on a regular basis. xDSL standards are relatively new and are, therefore, prone to change on a regular basis. This involves regular, programmable equipment up- grades such as software download of new versions of xDSL transmission and application code from a central management station.

DATA COMMUNICATIONS MANAGEMENT

As with all new technologies, technical and cost issues do get resolved in subsequent generations. xDSL is no different. The market demand for high-speed broadband communications will drive the development of the technology and will produce an economic, reliable solution for end users. For network operators, xDSL is the most cost-effective way of up- grading the copper infrastructure and competing against fiber and cable competitors. Another factor in determining market direction is the move for European PTTs (Post, Telegraph, and Telephones) from monopoly to deregulation and free competition. This will force network operators to unbundle their loops in a similar way to what has happened in the Unit- ed Kingdom and the United States. These new competitors will aggres- sively target end customers with offers of high-speed, broadband data communications, tempting them from the traditional service providers who may not be moving as quickly toward new service provision. So, using the existing copper networks, xDSL will be made available — which will motivate PTTs to respond, leading to a critical mass and ex- plosion in xDSL deployments. While network operators, service providers, and vendors debate the standards, technology, and cost issues of deploying xDSL networks, customer demand is growing for cost-effective, high-speed broadband networks that unleash new services and new business potential for them. Looking at the rapid development of Inter- net and intranet business and services over the last few years, one can expect resolutions to the main issues and mass xDSL deployment within the next three to five years. On the other hand, cable modems will dominate the North American residential Internet access market by 2003, outpacing xDSLs lines 5-to-1, according to a new market study. More than 8 million cable modems will be installed by then, according to Forward Concepts, Inc., a Tempe, Ari- zona-based consulting and market research firm. By 2003, the base for worldwide residential broadband access will be close to 30 million users, but that is not enough to support everyone who wants a piece of the broadband market. The big losers in the push toward high-speed broadband data services will be the telephone enterprises if they do not depend less on lawyers and more on technology. Finally, some see the battle as being tipped in telephone enterprises’ favor because customers will demand the dependable service they are accustomed to having with their phones. Still, cable modem enterprises are, by default, winning the Internet access war among consumers because the phone enterprises have really not been as aggressive as they could be in offering high-speed broadband data services. However, it is not too late for phone enterprises if they jump into the broadband market soon. Nevertheless, they will be shut out if they are not into volume shipments of xDSL by the end of the year 2000.

CABLE VERSUS DSL

CONCLUSION AND SUMMARY In the race to provide high-speed data access to Internet users, DSL will win over cable modems. That is because copper networks already are entrenched, and enterprises will want to leverage their existing invest- ments in copper. The tug-of-war between DSL and cable modems is a closely watched battle, and nobody knows for sure which side will prevail. The predic- tion of a DSL victory has less to do with the merits of DSL and more to do with the existing copper infrastructure. Nobody has love or hate for DSL. They do not care; but it must be copper because one cannot dig it up. What has been achieved in elec- tronics for 30 years in communications has been to make copper more efﬁcient. So, DSL is just an extension of a 30-year trend that will not stop. It is too expensive to change the lines. As one gets critical mass behind some of the vendors, DSL wins for sure. DSL carries data at high speeds over standard copper telephone lines and allows users to surf the Net and talk on the phone at the same time, making it especially attractive for the home and small-ofﬁce workplaces. Finally, cable modems, meanwhile, are gaining popularity, and some heavyweight investors are betting on cable enterprises to provide ex- panded services in the future. Thus, confusion over standards, including ADSL, DSL Lite, and VDSL, poses an obstacle, but those issues will re- solve themselves in the next six to 12 months for sure.

Notes 1. Digital subscriber lines carry data at high speeds over standard copper telephone wires. With DSL, data can be delivered at a rate of 1.5 Mbps (around 30 times faster than through a 56 Kbps modem). Also, DSL users can receive voice and data simultaneously, so small ofﬁces can leave computers plugged into the Net without interrupting phone connections. Currently, DSL is expensive because specialized equipment — a splitter — needs to be installed at the subscriber’s location. DSL Lite, the consumer-ready version of DSL, requires no such splitter, and promises comparable access speeds at a cheaper rate. 2. The ITU, a Geneva, Switzerland-based international organization that governs the communications industry, recently approved the G.Lite standard, a lower-speed DSL technology aimed at the mass-market consumer. 3. DSL runs over existing phone lines, but it still allows one to make calls while online. 4. With either one, the $9 to $14 per month charge that Internet users ordinarily spend on an extra telephone line can be avoided. 5. Westell was founded in 1980 and is headquartered in Aurora, Illinois, west of Chicago.

Addresses 1. Rockwell International Corporation, 600 Anton Blvd., Suite 700, Costa Mesa, CA 92628-5090, 1999. 2. Cablevision Systems Corp., 1111 Stewart Avenue, Bethpage, NY 11714, 1999. 3. Nortel Networks Corporation, 8200 Dixie Rd., Suite 100, Brampton, Ontario, L6T 5P6, Canada, 1999. 4. Cable Television Laboratories, Inc., 400 Centennial Parkway, Louisville, CO 80027, 1999. 5. TeleChoice, Inc., 8555 N. 117th E. Ave., Suite #101, Owasso, OK 74055, 1999. 6. America Online, Inc., 22000 AOL Way, Dulles, VA 20166-9323, 1999. 7. Prodigy Communications Corporation, 44 S. Broadway, White Plains, NY 10601, 1999. 8. CompUSA, Inc., 14951 N. Dallas Pkwy, Dallas, TX 75240, 1999. 9. International Telecommunication Union (ITU), Place des Nations, CH-1211 Geneva 20, Switzerland, 1999. 10. University of New Hampshire, 4 Garrison Avenue, Durham, NH 03824, 1999. 11. Intel Corporation, 2200 Mission College Blvd., Santa Clara, CA 95052-8119, 1999. 12. Alcatel, 54, rue La Boetie, 75008 Paris, France, 1999. 13. Lucent Technologies, Inc., 600 Mountain Ave., Murray Hill, NJ 07974, 1999.

DATA COMMUNICATIONS MANAGEMENT

14. Compaq Computer Corporation, 20555 State Hwy. 249, Houston, TX 77070, 1999. 15. Dell Computer Corporation, One Dell Way, Round Rock, TX, 78682, 1999. 16. Bell Atlantic Corporation, 1095 Avenue of the Americas, New York, NY 10036, 1999. 17. GTE Corporation, 1255 Corporate Dr., Irving, TX 75038, 1999. 18. US WEST Headquarters, 1801 California Street, Denver, CO 80202, 1999. 19. AT&T Wireless Services, 7277 164th Ave. NE, Redmond, WA 98052, 1999. 20. AT&T Broadband & Internet Services, 9197 S. Peoria Ave., Englewood, CO 80112, 1999. 21. Excite@Home, 450 Broadway Street, Redwood City, CA 94063, 1999. 22. Cisco Systems, Inc., 170 W. Tasman Dr., San Jose, CA 95134, 1999. 23. Motorola, Inc., 1303 E. Algonquin Rd., Schaumburg, IL 60196, 1999. 24. Deutsche Telekom AG, Friedrich-Ebert-Allee 140, 53113 Bonn, Germany, 1999. 25. NorthPoint Communications, Inc., 222 Sutter Street, 7th Floor, San Francisco, CA 94108-4458, 1999. 26. Universal ADSL Working Group, The ADSL Forum Ofﬁce, 39355 California Street, Ste. 307, Fremont, CA 94538, 1999.

John Vacca is an information technology consultant and internationally known author based in Pomeroy, Ohio. Since 1982, John has authored of 27 books and more than 330 articles in the areas of Internet and intranet security, programming, systems development, rapid application development, multimedia, and the Internet. John was also a conﬁguration management specialist, computer specialist, and the computer security ofﬁcial for the NASA space station program (Freedom) and the International Space Station Program, from 1988 until his early retirement from NASA in 1995. John can be reached on the Internet at [email protected].