The cable industry says it's more serious than ever aboutserving the business market with voice and advanced data services. But after a year of preparations focused on this new opportunity and vendors taking longer than expected to get a new generation of advanced service modems in the pipeline, the question of when cable service providers will begin making a dent in the market is becoming an issue of growing industry concern.
"I don't feel good about that or confident that we're going to have that [access to standardized next-generation modems in the near term]," says Dan Liberatore, chief technology officer at Adelphia Communications Corp. (www.adelphia.net), a top-10 cable multiple system operator (MSO). But he speaks for many operators when he says his company isn't going to let the modem standardization process get in the way of its ramp up for business services.
"We see a huge opportunity for us in the small to medium-sized business markets," Liberatore says. If performing preliminary tests and market trials entails use of first-generation DOCSIS (data over cable service interface specification) modems, so be it, he says, noting that, for trial purposes, engineers can dedicate enough bandwidth to business services so as to avoid the congestion which the forthcoming DOCSIS 1.1 modems will be able to control under more normal, bandwidth-restricted conditions.
Comcast Corp. (www.comcast.com) evinces a similar attitude about advanced service trials, including packet voice and business services, which it has under way in various markets. "For commercial deployments we have to have DOCSIS 1.1 equipment, including the low-power versions of DOCSIS 1.1 modems," says Mark Coblitz, vice president of strategic planning at Comcast. "But I don't see anything standing in the way of moving through the test and market trial phases and on to commercial deployments when we're ready."
The commercial delivery of modems conforming to version 1.1 of the DOCSIS standard has been held up pending completion of the certification process at Cable Television Laboratories Inc. (www.cablelabs.com), which is the industry's way of assuring complete interoperability among products from multiple vendors. A year ago, when anticipation of the business opportunities surrounding DOCSIS 1.1 first motivated some companies to begin preparing for business service launches, it was thought that certification would be under way by the start of 2000, but the projected startup date for certification has been pushed back repeatedly ever since, with no modems having yet entered the certification pipeline at CableLabs as of late June.
"It's really out of our control," says CableLabs spokesman Mike Schwartz. "It's up to the suppliers."
The primary difference between DOCSIS 1.1 and the widely deployed version 1.0 modems has to do with how use of the hybrid fiber/coaxial (HFC) cable network return channel is controlled so as to assure guaranteed throughput for voice and mission-critical business data applications. For example, if market demand saturates a downstream channel dedicated to high-speed data or packet voice services, the operator can allocate another channel to handle the overload. The upstream channel options are far more restricted, operating as they do within the usable portion of the 5 MHz to 50 MHz "subband" of the cable radio frequency (RF) spectrum. This usually adds up to only 12 MHz or so worth of bandwidth, some of which must be dedicated to other purposes beyond transport of customer data transmissions.
While the specifications for DOCSIS 1.1 were completed last year, ongoing efforts among vendors to improve on the basic platform through use of software innovations that often depend on a particular vendor's interpretation of the specifications in conflict with other vendors' interpretations have left the process in a state of flux. Moreover, to accommodate a late addition to the standard known as "fragmentation," some vendors had to redesign their chips, putting them at a disadvantage against those who didn't.
Fragmentation, which segments packets into shorter pieces so as to fill all available bandwidth openings as efficiently as possible, is one of several innovations that is meant to improve operators' ability to use the upstream portion of the cable plant to support the QoS requirements of IP telephony and mission-critical business services. Other software-based elements include dynamic QoS, which supports a mixture of traffic flow categories through a single interface, and new scheduling algorithms that are designed to reduce latency and jitter and to make more efficient use of bandwidth.
Another element of 1.1, which was added relatively late in the spec-setting process, is known as "payload header suppression." This technique saves bandwidth in the downstream as well as upstream packet flows by deleting any packet header information between the modem and headend CMTS (cable modem termination system) that isn't needed for that leg of the IP signal transport.
"The technique adds a couple of bytes of information in the header to deliver enough information so that whatever is needed [out of what has been removed] can be restored at the end points," explains Mark Sumner, a consulting engineer with Motorola Corp.'s (www.motorola.com) Internet Networking Group. "You're taking away 34 bytes on average and adding two for a net gain in efficiency of 32."
The 1.1 specs also include a new layer of security that uses digital certificates for modem authentication. Version 1.1 also defines the approach for implementing IP multicast over cable networks.
Reaching Out
Despite the headaches surrounding commercialization of DOCSIS 1.1, momentum continues to build on the cable business services front. In fact, with the integration of voice and data and the support for guaranteed levels of QoS that come with version 1.1 of the DOCSIS modems, cable companies are experimenting with fixed wireless and new optical solutions that would work with the DOCSIS modem platform to extend cable services' reach beyond the segments of the business market that are passed by existing cable plant.
For example, a unit of Adelphia that holds LMDS licenses across the country is working with Lucent Technologies Inc. (www.lucent.com) and Ensemble Communications Inc. (www.ensemblecom.com) to test next-generation broadband wireless systems for rollouts at as-yet-unspecified dates. And Comcast is testing a new low-cost means of extending fiber-based transport to commercial customers supplied by Quantum Bridge Inc. (www.quantumbridge.com).
Moreover, most major cable MSOs are affiliated with or fully owned by CLECs, including AT&T Corp. (www.att.com) with the former Teleport Communications Group Inc., Time Warner (www.timewarner.com) with Time Warner Telecommunications and Cablevision Systems Corp. (www.cablevision.com) with its Lightpath division. "With our licenses in wireless, a CLEC presence and cable coax, we don't care how we get to the customer," Liberatore says. "Our story is we're trying to reach as large a segment of the consumer and business markets as we can."
And, even without the noncable adjuncts to its operations, Adelphia and other cable companies are well positioned to reach the lion's share of the targeted business market, Liberatore adds. "Our research shows that between six and seven out of 10 small businesses are within 200 feet of our existing [cable] infrastructure," he says.
Cable ISPs
With two-way activation of more than half of the nation's cable infrastructure now achieved according to industry analysts and engineering reports, the industry momentum behind entry into the business markets has reached the point where all three major providers of high-speed Internet access over cable have set up units devoted to making entry into this new arena easier for operators than it might otherwise be.
The latest cable ISP to move in this direction is Road Runner (www.roadrunner.com). The cable data service provider, a joint venture of Time Warner, MediaOne (www.mediaone.com), Microsoft Corp. (www.microsoft.com), Compaq Computer Corp. (www.compaq.com) and Advance/Newhouse (www.timewarner.com), has put together a staff headed by former Ernst & Young (www.ey.com) senior manager Mark Mercer to market a new "Road Runner Business Class" suite of e-business solutions.
"Road Runner has been trying to get there for the past year, but the timing wasn't right until now," Mercer says. "Now we believe there's an opportunity to take what we've learned working with affiliates who have been offering business services and apply that knowledge as a new business unit within Road Runner."
Those affiliates, including Time Warner systems in Columbus, Ohio; Portland, Maine; San Diego; and Tampa Bay, Fla., have now amassed in excess of 10,000 commercial service customers, ranging from single users in the small office/home office (SOHO) market to small businesses to a handful of big corporations who are using Road Runner to deliver telecommuting services to workers at home, Mercer says.
"The target market for our services will be businesses with 100 or fewer employees and the enterprise telecommuting sector, where we can use the HFC infrastructure to reach end users," Mercer adds.
Road Runner and its affiliates who are already providing business services have learned how to market effectively to businesses that are within easy reach of cable plant, avoiding the problem of overburdening themselves with demand from businesses that are too far from the network to be served, Mercer notes. The company also can bring its experience at dealing with value-added resellers to the cable market, assisting operators in linking up with the experts who small businesses rely on to provide technical assistance with internal networks and equipment beyond the point of Road Runner's "demarcation," which is the cable modem, he adds.
Mercer draws a distinction between Road Runner's effort and that of the @Work unit of Excite@Home Corp. (www.corp.excite.com), which has been offering commercial services for more than two years, primarily over DSL and other noncable facilities. "They're using the @Home backbone infrastructure but have not been focused on cable, whereas we're looking at the opportunity to use cable infrastructure where it passes office buildings, campuses and strip malls to provide commercial services," he says.
However, while Excite@Home's @Work unit has built a customer base of typically larger customers than those courted by Road Runner's affiliates via alternative access modes, it is once again focusing efforts on cable after relatively futile attempts during its early operational phases.
"We're definitely seeing a big increase on the cable side of our business," says Sue Vaillarcourt, spokeswoman for @Work. "We expect cable to be a big piece of our revenue stream this year."
One of the basic forces behind this shift is the market awareness of high-speed access and the resulting customer-driven demand for connections from the small-business sector. "Public awareness of the @Home service has generated a lot of calls from businesses for service," Vaillarcourt says.
While Road Runner is starting out with a relatively small menu of options, including basic high-speed "best effort" Internet access and e-mail, it will move quickly to add more options, starting with hosted e-commerce service, which already is undergoing tests in four markets and will be expanded to half a dozen more in the months ahead, Mercer says. This quarter the company expects to add more sophisticated elements to the menu, including managed firewalls, VPNs and data storage.
"In the fourth quarter, we're looking at beginning to do some hosted application services," Mercer says. "We're in discussions with various players in this space, where we think there's a huge opportunity to be a point of aggregation for ASPs."
ASPs
ASPs represent a fast-rising category within the ISP community where software applications such as Microsoft Office, inventory management and human resources administration support are provided from off-site servers over high-speed links, alleviating the pressure on corporations' in-house information technology personnel to manage and upgrade an ever-expanding array of software packages. "We believe the utility and gateway that Road Runner provides for business-oriented content also can become a portal to aggregated ASP services," Mercer says.
The role of ASPs as a potential revenue stream for cable has yet to take shape in mainstream cable thinking, but it's clear that, with opening of the HFC-based data service offering to businesses, this could change quickly. For example, AT&T has committed $250 million of current infrastructure investments to develop what it calls an "ASP ecosystem" that will provide support for intelligent content distribution and managed storage services, advanced caching and backup for server failures, high-speed backbone transport and local access distribution via DSL, now available in 50 local markets from AT&T nationwide.
Sources say it's just a matter of time before the carrier's MSO arm opens its HFC pipes to broaden the base of users ASPs can reach via the AT&T infrastructure. "It won't be long before this becomes a no-brainer, but we're not quite there yet," says an official working on business applications at AT&T Broadband and Internet Services.
G.lite Picture Brightens
Standards, Line Sharing Spotlight DSL Deployment
By Gail Lawyer
G.lite could be considered the Miller Lite of DSL technology. The ability to get broadband Internet access at low prices leaves a great taste in consumers' mouths. And the fact that G.lite can be self-installed makes it less filling for service providers that previously had to send a technician out to install every line.
But, after a year or so of promises that availability would be widespread in 2000, G.lite deployment is just getting under way. Delays caused by standards development and the necessary interoperability testing for CPE and CO equipment, pushed G.lite deployment back farther than many industry watchers anticipated.
Now, though, the standards have been set. And, with the introduction of line sharing in early June, enabling CLECs to more easily provide DSL to consumers, G.lite is poised to take off.
With more widespread usage of G.lite, the industry can expect to see quite a jump in DSL deployment. For instance, TeleChoice Inc. (www.telechoice.com) predicts that between 2000 and 2001, the number of DSL lines in service in the United States will jump from 2.1 million to 5.1 million, a 142 percent increase. A large part of that growth can be attributed to G.lite, believes Mike Crawford, director of new product development at GTE Corp. (www.gte.com).
What exactly is G.lite? G.lite is a type of asymmetric DSL (ADSL). ADSL is the most widely used version of DSL to provide broadband services to residential customers. It's ideal for delivering Internet access because it reserves more bandwidth going downstream to the user than upstream to the Internet.
There are several flavors of ADSL, including G.dmt and G.lite. G.dmt, also referred to as full-rate ADSL, can provide up to 8mbps downstream and 1.5mbps upstream. To use this type of DSL, the service provider most likely will have to install a splitter on the phone line in the home, requiring a truck roll.
G.lite, though, is what many consider to be "universal" ADSL, a new standard for DSL that was adopted in mid-1999. G.lite is universal because it can be installed by the end user, rather than requiring the service provider to send out a technician. Cost for the equipment and the service will be less than other ADSL varieties because chip sets used in the modems are less expensive and G.lite speeds max out at 1.5mbps downstream and 512kbps upstream.
Many service providers were expecting that 2000 would be the year of G.lite. But the year didn't quite start out that way.
"We had planned it for 2000, but we're not there yet," says GTE's Crawford. "G.lite deployment schedules have slipped from what we expected a year ago."
Some telcos, like GTE, have been using other types of proprietary splitterless DSL technology that allows them to offer customers the self-install option. But rather than using ATM for aggregation, as is done with the true G.lite model, GTE used frame relay because it was already in place and available, says Crawford. As of late June, GTE had well over 120,000 DSL customers, with about 80 percent of them installing DSL themselves.
However, GTE is not giving up on deployment of true G.lite. "GTE is intensely interested in G.lite, and we're active in all the forums," says Crawford. "We believe G.lite is viable going forward, although there will be more [other] options than we originally expected."
For CLECs, it was more than just standards and equipment interoperability causing delays in G.lite deployment. There was also the issue of access to the incumbents' copper loops. Before incumbents were required to share the loops, additional lines had to be installed for a competitor to provide DSL service to consumers or businesses.
But, with the mid-2000 advent of line sharing, CLECs now are able to offer data services over the same line on which incumbent telcos are providing voice services. And, because no new line must be installed, truck rolls can virtually be eliminated.
"We've been waiting for line sharing to happen for G.lite. It's a beautiful merging of these two things," says David Angell, DSL evangelist for NorthPoint Communications Group Inc. (www.northpoint.net) and author of the book DSL for Dummies. "Line sharing is the driving force to get it deployed. And it will allow CLECs to offer consumer DSL products at a cost some fraction of what it would be for a whole line."
NorthPoint has completed its trials of G.lite and now is in the process of deploying it throughout the U.S. markets it serves. NorthPoint plans to use G.lite in its launch of consumer services.
"The consumer space didn't have competition before. CLECs have primarily been doing competition in SDSL [symmetric DSL] for business and weren't in residential because they couldn't compete with the BOCs," Angell says. "With G.lite, it's about shopping around for DSL service. And all of a sudden, CLECs are serious contenders in the consumer DSL space."
The reason the combination of G.lite and line sharing are making consumer offerings from CLECs possible is the ease of delivery G.lite promises. Because customers can install G.lite themselves, truck rolls will not be necessary for most lines. CLECs can ramp up the deployment of G.lite-based DSL service without having to worry about hiring several hundred new technicians and customer support staff members.
"With line sharing, customers can do self-installation. We don't have to do anything with the inside wiring. They just plug it into the jack," says Mike Calabrese, vice president of product management at Rhythms NetConnections Inc. (www.rhythms.net).
Rhythms announced in early June that it is implementing service across the country over shared lines via G.lite and G.dmt. Rhythms has line-sharing agreements in place with Bell Atlantic Corp. (www.bell-atl.com), BellSouth Corp. (www.bellsouth.com), GTE, SBC Communications Inc. (www.sbc.com) and US WEST Inc. (www.uswest.com).
While G.lite is an ideal technology to bring less costly broadband to the masses, its use poses some network questions for service providers. With ever-increasingly complex websites, as well as voice and video over DSL, consumers are going to continue to demand faster and faster connections to take advantage of new web services. But service providers don't want to have to upgrade their networks to handle higher speeds after just putting in equipment that is meant for G.lite access.
So, rather than just build for G.lite, service providers are selecting CO equipment that can handle any type of modem an end user might have.
"Most [service providers] look at DSL deployment from a CO perspective," says Jay Fausch, senior director of marketing for the DSL business unit at Alcatel USA (www.usa.alcatel.com). "They don't want to have to reinstall full-rate ADSL when they can throttle that down in 32kbps increments."
Rhythms, for example, is using DSLAMs from Cisco Systems Inc. (www.cisco.com) that include software-selectable DSL features. Rather than having to install separate equipment for G.lite and G.dmt, Rhythms is able to use a single DSLAM and provision either DSL flavor through the equipment's software. "We very much believe it's an advantage that we can do both on a single platform," says Calabrese. "G.lite is for the low-end consumer, but we believe that G.dmt will eclipse G.lite."
Covad Communications Co. (www.covad.com), too, upgraded its network in March so that both G.lite and G.dmt are supported, says Eric Moyer, Covad's director of product marketing. Covad's consumer offering--TeleSurfer ADSL and TeleSurfer Pro ADSL--are based on G.lite.
While G.lite deployment has begun, the promise of DSL modems being available in retail stores is still relatively nonexistent. There have been a few announcements in specific markets in which service providers are working with electronics retailers, such as Best Buy and Radio Shack, to stock DSL modems. But the average consumer can't just run out to one of these stores anywhere in the United States, choose from a variety of customer premises equipment, bring it home, plug it in and have DSL.
"Today's retail market is immature," says Calabrese. "The consumer doesn't care what [product] they get. They're worried about the cost and how fast they can get it."
In the current model, the service provider handles modem distribution, and the consumers have little choice in what product they get.
"Retail would be a big relief for the CLECs, and it's a key ingredient in mass deployment," says Dylan Larson, director of product development at New Edge Networks (www.newedgenetworks.com), a wholesale provider of DSL in small, medium-sized and semirural markets. The retail model "will relax some logistics for us. We've been able to keep up with our stock, but the fewer steps we have to take, the faster we can ramp up services."
