Carrying the standards

Apr 9, 2001 12:00 PM, Jim Barthold

What the heck is wrong with DSL? How can a technology that uses existing infrastructure to feed a ravenous customer base be taking a back seat to cable modems?

DSL is partially hindered by a fractured standards process, but standards — or a lack thereof — are not the crux of what's bothering DSL. Or so some might say.

“There are no standards issues blocking deployment of DSL,” says Greg Wetzel, who, when not serving as Covad's chief architect, chairs the DSL Forum's voice-over-DSL working group and the ATM Forum's architecture working group.

“As far as the basic standards for ADSL are concerned, they were done some years ago,” says Martin Taylor, chief technology officer for CopperCom, confirming a vendor's point of view. “Interoperability is now just totally taken for granted as far as basic DSL is concerned.”

Of course the industry is plagued by finances — or again, a lack thereof — but that's probably only a natural progression and shakeout of people with money who realize that you don't get more money if you keep giving it away.

“It's not news to anybody that DSL, from the financial analyst basis, seems to be the industry to beat right now,” says Keith Markley, president and chief operating officer of DSL.net. “We suspect that will only last so long. As long as users are buying the end product and paying more for it — and the base is growing — at some point that will no longer be the popular song to sing.”

Finally, DSL is designed for something else and is being forced into the square peg space. DSL was invented to deliver video over twisted pair. Data happened when the Internet got hot and TV cooled off.

Whatever the case, DSL is in the back seat of a high-speed vehicle being driven by cable modems, and it doesn't look like it will be taking control any time soon. Kinetic Strategies predicts that cable modems will not only keep their market lead but will expand it going into 2005, based primarily on the multiple technological and operational difficulties facing DSL.

“There is a myth that DSL is a technology that allows RBOCs to effortlessly use their legacy copper plant to offer broadband. That turns out to be false,” says Kinetic's president Michael Harris. “It requires quite a bit of reworking.”

So what's stopping that rework?

The best efforts

“Presently, standardization is not anything that's having any kind of major negative effect on the rollout of the technology or the services,” Markley says.

That's not to say that there aren't a bundle of confusing ongoing standards efforts. There's the DSL Forum, the ATM Forum, The Institute of Electrical and Electronics Engineers and any one of its multiple iterations, the International Telecommunication Union and the American National Standards Institute.

Each is working on a piece of the puzzle — sometimes at the same time in different places. For the most part, those pieces are minutiae. Two ongoing efforts, auto-provisioning and the symmetric delivery standard g.SHDSL, are being worked out.

“A lot of the issues we have with non-standard DSL deployment are going to go away with the deployment of g.SHDSL,” says Kumar Shah, chief marketing officer for AccessLan Communications.

Right now, most providers embrace an asymmetrical DSL (ADSL) delivery method that's better for residential users.

“Business users typically want voice, and they may be hosting Web sites and that sort of thing and typically want a more symmetrical DSL arrangement,” says CopperCom's Taylor. Therefore, CopperCom offers a symmetrical architecture (Figure 1).

The industry is also not holding its breath waiting for autoconfiguration, even though that will ease the pain of installation.

“The autoconfiguration stuff is still an evolving standard but the cooperation of the industry is very much into it and working diligently and will have probably a proposal out this year,” says Jay Fausch, senior director of marketing of Alcatel's DSL business.

The DSL Forum is pursuing autoconfiguration and hopes to have it resolved soon. “The work is an effort to define an open standards-based way in which anybody's modem can be attached to anybody's DSLAM and that the communication that occurs between those devices automates those configuration steps,” Fausch says.

That's a big step in subscriber provisioning, but again, it's an addendum to what's already there; a lot of the pre-provisioning hassle is already under control.

“In most cases, the equipment you use for testing in the central offices [provides] a pretty good idea that it's going to work,” Markley says.

Standards, he says, don't help or hinder the installation process.

“With all these things you have a pretty good idea. Subsequently, the final test is when you install,” he says.

The business case

Money markets have a hard time swallowing DSL's business model of upfront investment to gain recurring dividends.

“A dotcom will spend $1000 to acquire an eyeball [and] that eyeball has to hit their site over and over again for them to recover revenue,” says Michael Malaga, founder and former chairman of service provider NorthPoint Communications. “In our business, we spend our $1000 or whatever it takes to acquire the end user and they pay us a recurring revenue stream.” [Since this article was written, NorthPoint has gone bankrupt and sold most of its assets to AT&T. (Telephony, April 2, page 26)]

That model parallels cable, but cable is using its data over cable service interface specifications (DOCSIS) effort to drive down prices and push products into retail.

‘There are no standards issues blocking deployment of DSL.’
— Greg Wetzel, Covad

“We didn't just create DOCSIS. We created the whole process,” says Rouzbeh Yassini, CEO of Yas and an executive consultant with CableLabs. “We have an end-to-end IP platform strategy that all rides on top of DOCSIS… or runs on top of DOCSIS' achievements by getting the vendors and market engaged. Cable is doing a fabulous job.”

DSL supporters, of course, won't call cable's approach superior.

“We don't have a standard like DOCSIS,” Wetzel says. “I think the DSL Forum specifications may be more flexible than DOCSIS. There are fewer restrictions on the service provider and the way things can be arranged.”

Cable starts with a built-in or built-out advantage, says Michael Goodman, a senior analyst with The Yankee Group. “You have a brand new plant when you're talking about cable modem service because they've had to upgrade the entire thing.”

DSL's installation scenario is less assured, he says. “Because there is no guarantee until somebody has actually come out and determined whether or not you have a clean circuit that they can use. Installing cable modem is more science than art. Installing DSL is more art than science. And it's just a lot more complex.”

The cellular model

DSL has been modeled more closely on the mobile wireless industry than POTS or even cable television. In cellular, it's not that important what standard is being used. “If I want cellular phone service today, I go to a service provider who has a number of different phones,” Fausch says. “I select one of those and they may give it to me or sell it to me in return for me signing an agreement to do business with them. That's worked extremely well for the cellular industry, and I see no reason why that could not work in a broadband service model.”

In that kind of model, standards are just not that important, adds Malaga. “We can use any one of a number of different standards or any one of a number of different flavors of DSL in any one of our central offices, including standards-based G.Lite.”

The biggest challenge, he says, is getting to the subscribers. “The LECs have done more to impede the growth of this industry than anything,” he says.

Line sharing is supposed to change that.

By sharing the same copper, the data LECs should move to equal footing with the incumbent carriers.

Should.

“We have committed to 100% line share implementations for consumer lines going forward, and that has a significant positive effect on the costs and simplicity of installing DSL,” says Chuck McMinn, Covad's chairman. “The end CPE and interoperability is really not the driving factor in how fast DSL rolls out. These are not technology issues; they're operational issues with telephone companies doing their work.”

The dark side of line sharing is that it encourages the continued use of DSL, a technology that squeezes the last ounce of capability from old copper networks. Verizon had an aggressive fiber deployment strategy that hit a significant pothole when DSL hit the radar screens.

“In a lot of ways DSL has slowed down fiber deployment,” says John White, Verizon's wholesale technology executive director. “We've… ended up guaranteeing CLECs that if they buy a loop that's on copper, we'll keep it on copper. Now we're locked into an all-copper base.”

Migrating from fiber is good for DSL but bad for the evolution of future forward-looking networks, he says.

Short-term line sharing helps the ILEC/CLEC relationship, White says. “The CLECs have built and turned up the line sharing and they should do really well with it.”

Malaga disagrees about how cooperative the “resistant-to-change monopoly provider that holds the keys to the infrastructure” is being.

Chalan Aras, vice president of advanced technology and product planning for Jetstream, agrees that the CLEC/ILEC war isn't helping DSL. “Part of the problem is the ILEC is clearly not motivated to give fast service to the CLEC,” he says.

Cable's advantage, for now at least, is that it doesn't have to share its networks.

“If the laws change tomorrow and say cable shall give equal access to anyone, then you would have the same exact problem,” Aras notes.

The incumbents are slowing the competition in the DSL space, says Frank Gangi, president/CEO of Global NAPs, which has built an East Coast network that it plans to expand nationally (Figure 2).

“The real issue is getting the Bells to put the wires up from the customer site to your DSLAM. You need the life span of Methuselah to deal with these guys,” Gangi says. “That's the biggest problem with DSL: the dependence on the regional Bells.”

The regulators

White wants Verizon's networks to offer more capability — meaning fiber — and build bandwidth and provisioning comparable to what the competition brings over hybrid fiber/coax networks.

“We should be getting the fiber out there, but I have to tell you we're in a regulatory quandary,” he says.

Regulation — or more exactly, lack of it — is what most DSL providers claim gives cable its edge.

“They don't have to unbundle. They don't have to sell it at a retail price to somebody else. They can make money at it,” says White. “We can do fiber-to-the-curb, we can do fiber-to-the-home, we can give a better package of services and compete with them, but we're restricted from competing because if we make that investment we have to unbundle it, sell it at prices that do not recover our costs. That's what's going to hurt us from getting true competition in IP access straight to the customer.”

At a store near you

DSL is not as advanced as cable in the retail marketplace. You now have a chance of getting a cable modem at the local electronics store; you'd be hard pressed to find a DSL modem on nearby shelves.

“I definitely think standards have a lot to do with it,” says Charlie Twomey, Genuity's senior manager of DSL product management. “You'd want to be able to walk into a Best Buy or Circuit City anywhere in the country and buy DSL, take your computer home and plug-and-play. Certain companies are making moves to try to get their products and networks and pricing and technologies more standardized, but I still think we're a ways away from being there.”

The roadblock to consumer deployment in large scale was the line sharing, says Wetzel. “Soon you will be able to buy ADSL CPE in certain stores.”

And then what happens once it reaches the home?

“There's a need to educate the general consumer on technologies like HomePNA in which you would enable every phone jack in your home to be a network connection,” says Morgan Chase, senior systems solution manager of Nokia Broadband Systems. “The central issue is once I have that broadband connection to my home, how do I then circulate it through the multiple rooms in my home?” Yassini says.

Cable again has an edge.

“DOCSIS has become almost a de facto interface standard where all manufacturers from the computer side to set-tops to Samsung refrigerators all utilize that type of thing,” Yassini says.

It starts, of course, with the cable modem. “Back in 1988 a cable modem was $15,000. In 1995 it was around $495. Today you're averaging around $100, and by next year it will be well below that, in the two-digit area.”

And where's DSL? Still in the proprietary model.

“We supply the modem to the person at the other end, and we have the equipment in the central office. As long as we can control both ends of that piece of copper, it doesn't matter what standards are, really. Standards help, but they certainly don't hinder,” says NorthPoint's Malaga.

DSL.net's Markley disagrees: “The best thing about standards is that they create competition in the equipment marketplace and drive commonality in people's homes, so eventually you start to drive the cost of CPE down and your end user education gets better.”

Winning friends

In the end, the DSL dilemma has little to do with technology and less to do with standards and most to do with the public perception of the service as proprietary, difficult to get and expensive to start up.

“It really isn't proprietary in any way,” says Alcatel's Fausch. “The people who are building DSL products today are adhering to an industry specification that is an open standard. The autoconfiguration stuff is still an evolving standard, but the cooperation of the industry is very much into and working diligently.”

“It's a lot more complex because you have more players involved,” says Goodman. “You have to line up the phone company. If you're using a reseller… you have to get them on board and then you have to get the ISP on board. Then you have multiple customer service departments. Multiple calls by technicians and multiple billing and customer care issues that have to be resolved.”

Then there are just the fundamental flaws.

“The whole DLEC [concept] was flawed from the first and it just took investors a while to savvy up to the reality of trying to sell RBOC network and claiming yourself to be any kind of carrier,” says Kinetic Strategies' Harris.

And customers want the service.

‘Standards help, but they certainly don't hinder.’
— Michael Malaga, NorthPoint

“The only problem is that the financial markets have dried up. We're a capital-intensive business. The demand for DSL is there,” says Covad's McMinn.

Perhaps too much demand, which is leading to the perception that DSL — for whatever reason — can't deliver.

Those are perceptions fueled by “analysts and people who are pundits of the industry, as opposed to the market that we're selling the product into,” says DSL.net's Markley. “Any time you have a situation where the end user base continues to seek the product, it's hard to accept the statement that you're losing the perception battle.”

It should, then, be even harder to accept that DSL is losing the battle to cable — an industry that is hardly renowned for its ability to work together and for whom customer service frequently translates to lip service and public relations-inspired slogans.

“The whole issue between DSL and cable boils down to nothing more than time,” says NorthPoint's Malaga. “Cable has had just a lot more time.”

Really? Not when it comes to standards, according to that same Malaga.

“DSL standards are farther along than DOCSIS is, to put it bluntly. ADSL and G.Lite really were standardized last year and DOCSIS only started producing equipment in volume in the past quarters.”

And yet, perceptions being what they are….

“Our view at The Yankee Group is, and has been, that cable modem will stay ahead of DSL,” says Goodman.

Even its supporters concede DSL's flaws. “It's more difficult to install a DSL modem than a cable modem right now, but at the end of the day the number of installed DSL lines is going to exceed the number of cable modem lines, despite the fact that the cable modem industry has had a three-year head start,” says Covad's McMinn. “Somehow the DSL industry is getting it done and getting it done at a pace faster than the cable industry is.”

With or without standards.

“Most of the customers don't care whether you're using standards-based ADSL or proprietary SDSL or what have you. It's more about the implementations and the value of high-speed access,” says Greg Langdon, executive vice president of product strategy for Efficient Networks.

And who gets into the home first.