Why Canada's cable telephony trials crashed and burned
Though ambitious, Cogeco’s and Videotron’s initial work with packetized voice services was fraught with technical and operational tribulations
Just a few short years ago, Voice-over-IP (VoIP) via a cable TV infrastructure seemed a sure bet in Canada. In fact, both Groupe Videotron and Cogeco Cable–at the time Canada's second- and fourth-largest MSOs respectively, in terms of number of subscribers–were diving headlong into VoIP.
To say the least, they were bullish about VoIP's chances. For instance, Cogeco President and CEO Louis Audet told his shareholders in 1998 that VoIP could be offered "at a capital cost lower than $600 (U.S. $400) per telephone customer. This is about half of the $1,200 (U.S. $800) associated with traditional technologies." His goal: to launch commercial VoIP service through Cogeco's cable modems. Cogeco's networks were based on DOCSIS 1.0 specifications, but plans were in place to upgrade to 1.1–and the quality-of-service it brought with it for packet voice services–within a year.
At roughly the same time, Groupe Videotron's founder and then-Chairman Andre Chagnon predicted that, by 2003, half of his company's revenues would come from local and long distance voice combined with residential and commercial high-speed data services.
A year later, Groupe Videotron's then-VP of IP Telephony Francois Laflamme said his company would "be putting together an overall business readiness test that will enable us to get to service next year (2000)." His goal: for Groupe Videotron to offer commercial VoIP to two million households in 2001.
However, as the old saying goes, "Man proposes; God disposes." By 2002, Canadian VoIP was about as lively as Excite@Home. In other words, quite dead. All the bold pronouncements had come to naught, along with the careers of many who made them.
So what went wrong, and what can be learned from Canada's abortive VoIP cable trials?THE LATE 1990s: SETTING THE STAGE
It's almost hard to believe, but the cable TV world was a very different place a few short years ago.
High-speed Internet access was starting to take off, and MSOs who had traditionally survived on video alone were suddenly provided with a new revenue stream. Meanwhile, digital cable was inching up on the horizon–albeit later than everyone had expected it to–and "convergence" was the word on everyone's lips.
In such a heady environment–for heady it was–adding VoIP to an MSO's bag of tricks seemed to make sense. After all, conventional wisdom said VoIP is nothing but data packets by another name. How hard could it be to add them to an MSO's shiny new two-way plant? Actually, it turned out to be difficult, as most acknowledged back in the late 1990s. That's why it was left to Groupe Videotron and Cogeco Cable to blaze the VoIP trail in North America.
In each case, it was the character of the families in charge that drove their VoIP gambles.
Over at Quebec's Groupe Videotron, Andre Chagnon had made his fortune through technological innovations like Videoway, the primitive two-way analog terminal which nevertheless boasted 250,000 users by 1999. He innovated out of necessity: lacking enough French language channels to attract the viewers he required, Chagnon knew he had to market Groupe Videotron by other means.
Meanwhile, Cogeco Cable's Audet family also had a strong history of technological risk-taking. A case-in-point was Cogeco Cable's pioneering role in deploying DOCSIS 1.0 modems in the late 1990s; months ahead of many other MSOs. In fact, Cogeco Cable's deployment was so advanced by 2001, that by the time Excite@Home's backbone shut down in late February 2002, Cogeco Cable already had one of its own.THE TRIALS
When it came to rolling out VoIP, Groupe Videotron's plan was simple– first start with a trial involving about 200 homes in the Montreal area, then expand to 2,000. When the bugs had been ironed out, roll out to the entire province of Quebec, and that was that.
Cogeco Cable's tactics were somewhat different. According to the May 2001 edition of the Canadian trade magazine Cablecaster, the MSO wanted to target market VoIP to its 90,000 cable modem subscribers.
The company's plan was to turn the province of Ontario into a "big local area," said Dan Delisle, cited in the article as Cogeco Cable's senior business consultant for VoIP. As for pricing, Delisle said, Cogeco Cable hoped to make VoIP a flat-rate addition to Cogeco@Home, rather than charging via traditional local/long distance tariffs.
In preparation for that rollout, Cogeco Cable licensed itself as a CLEC (Competitive Local Exchange Carrier) by the Canadian Radio-television and Telecommunications Commission. As for the nuts and bolts of deployment? "Basically, we started this endeavor in mid-'99, and actually installed the first few customers by the end of the year," Francois Audet, Cogeco Cable's director of telecommunications, tells CED.THE TECHNOLOGY
To make VoIP happen, Groupe Videotron turned to Cisco Systems and Telcordia Technologies (formerly Bellcore).
Specifically, the MSO chose a product suite based on Cisco IP hardware. It included Cisco's uBR7246 Cable Modem Termination System, AS5300 Voice Gateways, 12000 Series gigabit switch routers, and Catalyst 8500 Multiservice Switch Routers.
Meanwhile, the software smarts were provided by Telcordia's Intelligent Gateway Call Server System (IGCS), and the cable modems were supplied by Samsung.
"Cisco technologies integrate extremely well with our existing infrastructure and allow us to bring Internet protocol data, voice and cable services to new, growing markets," said Claude Chagnon, then-president and chief executive officer of Videotron Communications Inc. (and son of Andre). "With the Cisco solution, we will be able to provide our customers with reliable, secure telephony services, highspeed Internet access, high-quality cable services and new emerging applications using the Internet Protocol."
Groupe Videotron officials were contacted for this article, but declined comment.
Over at Cogeco Cable, the company had already chosen Cisco as its primary high-speed Internet access vendor. Hence, Cogeco Cable was already equipped with Cisco uBR7246s with integrated MC16 modem cards. Expanding to VoIP simply meant adding extra Cisco and Telcordia products.
So how did two reasonably well-planned trials–backed by two engineering-inclined MSOs, and supported by Cisco and Telcordia–end in defeat?
There are two reasons: the first is technology; the second, economics.
First, the technology. Like so much in the early days of two-way cable plant, it just wasn't ready for prime time and wasn't necessarily outfitted for VoIP.
"While the technology was certainly getting close to deployability, we realized that there was still a significant investment required to make it work," says Francois Audet today. "We concluded that it was beyond our capability, as a smaller MSO, to underwrite the expense."
"We're too small to influence the development of that by ourselves," adds Denis Belanger, Cogeco Cable's VP of engineering. "It has to be an industry effort, and so far, it's not been an industry effort."
Without a doubt, scalability was a big challenge, says Belanger. However, he adds that there were also problems interfacing with ILEC switched networks, and providing 411/611/911 services.
These latter problems are news to Jonah Ninger, Telcordia's executive director. "Looking back at our records, Telcordia's Call Agent successfully passed the Bell Canada POI (point of interconnection) tests by November 30, 1999," he tells CED. "On December 5, 1999 we completed PSAP testing with five separate Montreal-based bureaus to validate 911."
Beyond the big problems mentioned by Belanger, small details caused headaches; the kind that only crop up during actual trials.
A case-in-point was the insertion of voice data packets into the DOCSIS network, says Mark Bakies, Cisco Systems' director of marketing. In its simplest terms, the VoIP technology of the day sometimes had trouble coordinating voice packets with data packets, he tells CED. This resulted in millisecond bursts where the network bandwidth was essentially empty, and added to latency delays.
As well, "the gateways weren't redundant enough; at least not for carrier-grade switching," Bakies says. In addition, "The softswitches could do the job, but they were more difficult to manage, and weren't always optimized in the right places."
Offering secondary-line VoIP services over DOCSIS 1.0 networks, however, has become a much more stable proposition. Time Warner Cable, for example, is now using what cable insiders call DOCSIS 1.0 "plus" to inject QoS capabilities into its initial LineRunner VoIP market trials in Portland, Maine and Rochester, N.Y. Importantly, DOCSIS 1.0 "plus" is not part of any official CableLabs efforts. The "plus" part refers to a proprietary QoS element that Cisco offers to MSOs that want to offer second-line VoIP services to DOCSIS 1.0 cable modem customers.
Now, the economics. As a result of the unforeseen teething pains of VoIP technology, deploying voice over cable no longer fit its original business case.
But this wasn't the only factor. "Local phone rates in Canada are very low," says Ian Angus, president of the Toronto-based telecom consulting firm Angus TeleManagement Group. "This makes competing against the ILECs tough on challengers; even without VoIP's teething pains. As well, one reason why Canada's MSOs wanted to move into local telephony was their fear that ILECs like Bell Canada would move into video and offer TV and telephone service as a bundle. That hasn't happened, at least not over telephone lines, so the move into telephony is less urgent."
The bottom line: No longer was VoIP cheaper to deploy than conventional telephony, as Louis Audet had hoped. Gone, too, was the fond hope of Andre Chagnon, that local/long distance telephony would be a major part of Groupe Videotron's bottom line by 2003.
In fact, gone, too, was Andre Chagnon. Forced into an unwanted takeover bid by publishing giant Quebecor, the Chagnon family was bought out for U.S. $3.4 billion in 2000. Then the dot-com bubble burst, and the debt Quebecor had incurred in taking over Videotron began to tell on the publisher's bottom line.
The result: although Groupe Videotron's VoIP service eventually signed up 2,000 customers, the project was killed by a cash-strapped Quebecor later that same year.
Meanwhile, Cogeco Cable's own trials–which had made it to 65 "friendly households"–were axed soon after.
Again, the reason was money: VoIP hadn't made any for Cogeco Cable. In fact, the MSO took a U.S. $20 million writeoff in 4Q 2000-2001, when it shut down its VoIP project.ARE WE THERE YET?
There's no doubt that MSOs want to deploy VoIP. After all, they're in the business of making money, and VoIP is one way to make it. Someday. "We remain very interested," says Francois Audet. "But what we want to see is hardware and software that you can plug-and-play. We want it to be ready for prime time."
So is VoIP ready for prime time today? Yes, says CableLabs CEO Richard Green. "I think the technology is ready for deployment," he declares. "I suspect, in this year, we'll see actual IP voice customers that are not just trial customers, but true business customers."
Telcordia's Jonah Ninger agrees. "The technology is ready for prime time," he says. "However, the MSOs' infrastructure is not universally upgraded to support the capacity and quality of service required to transport voice. Equally important are the operational and business pre-requisites, such as having telephony experience, and managing the regulatory requirements, and some MSOs are better prepared than others.
"We are excited about the success that we have had in Rochester, N.Y. with Time Warner," Ninger adds. "Having had 1,000 paying VoIP subscribers up for close to a year utilizing the Telcordia Call Agent and Cisco infrastructure, we look forward to expanding that into a commercial launch later this year."
For his part, Cisco Systems' Mark Bakies is more cautious about VoIP's prime time readiness.
"I think everybody's definition of ready is a little bit different," he says. For companies only wanting voice over data, or willing to help troubleshoot VoIP's remaining bugs, then yes, it is ready. But for MSOs who want plug-and-play, it isn't.
Until this happens, don't expect to see VoIP returning to Canada anytime soon. In addition, "The trend is similar in the U.S.," says Imran Khan, a senior analyst with the Yankee Group. "Cable companies here would rather not have any short-term revenues from voice, than hurt their brands by pushing 'work-in-progress' type of solutions to the end users."ONE BAD APPLE DOESN'T SPOIL THE WHOLE BARREL
Although Videotron's and Cogeco's ambitious trials became a tangled mess of uncooked economics and technology, the leap that the cable TV industry has made over the last few years is profound. CableLabs has started testing equipment based on PacketCable specifications, and many MSOs are embarking on or have already launched a variety of VoIP field trials–proof that their confidence in the technology is growing. Indeed, getting this far this fast is no small feat.
Still, not every project is fated to succeed. That's the price of innovation; they don't call it "the bleeding edge" for nothing.