Vendors put VOD on front burner
Spurred by aggressive moves by cable MSOs such as Time Warner Cable, Insight Communications, Cox and Comcast, the lucrative promise of video-on-demand (VOD) is finally morphing from a mere broadband pipe-dream to a budding reality.
Though 2000 has been witness to a surge of VOD deployments and market trials, VOD's true day in the sun won't likely arrive until next year. In fact, when it comes to what year VOD will make its first molar-rattling impact, the answer is virtually unanimous among equipment vendors, service providers and industry observers—they're betting on 2001, and betting on it heavily.
"We think 2001 is going to be the year for video-on-demand," forecasts Michael Poustovoi, associate analyst for research firm Morgan Keegan & Co. "That's when it will really ramp up and proliferate, not only in the U.S., but in the U.K. as well."
By the middle of 2001, a majority of the cable industry's heavy hitters are expected to be sitting dead red, prepared to knock the VOD ball out of the park. Adds Steve Nussrallah, president and CEO of video-on-demand player Concurrent Computer Corp., "I think by the middle to the end of next year, I would be surprised if you didn't hear announcements from six or seven of the top seven (MSOs) on VOD. I won't be surprised if the more aggressive MSOs forecast and budget for a significant portion of their systems to become VOD-active sometime next year."
There are several factors that are contributing to the VOD renaissance. For starters, streaming costs are not as staggering as they were five years ago. Gone are the days of cable operators paying $1,000 per stream. Advances in technology have caused those costs to plunge between $300 to $350 per stream, says Poustovoi.
According to Morgan Keegan, domestic video server equipment sales in calendar year 2000 should reach between $40 million and $50 million, and rise to $110 million to $130 million in 2001. But that's just the beginning. Within the next five years, conservative forecasts place that figure between $1 billion and $1.2 billion. More aggressive predictions put it as high as $2 billion.
While MSOs will inevitably drive the availability of VOD, they'll still have to lean on companies like Concurrent, Demand Video, DIVA, nCUBE and SeaChange to provide the technology that will aggregate and store the content.Concurrent Computer Corp.
From the outside looking in, Concurrent at first seemed like one of the most unlikely companies to join the video-on-demand brigade. More than three decades old, the Atlanta, Ga.-based company was known more for building flight simulators for NASA and Boeing, and manufacturing scanners that detect knots inside two-by-fours, than it was for making video servers and VOD software. Its real-time computer business realized the potential of VOD in 1997, the year Concurrent built its first server.
After that, Concurrent hawked its servers initially to markets it was the most familiar with, namely in-flight entertainment and corporate training. However, when the company wanted to market its wares to cable operators, it discovered that it lacked the industry experience it would require.
Enter Steve Nussrallah, a former Scientific-Atlanta (S-A) and Syntellect Inc. engineering executive who joined Concurrent as a consultant and board member in 1998. He was elevated to acting president of Concurrent's XSTREME division in 1999, and president and CEO of Concurrent early this year.
"When I came on board, Concurrent had a wonderful technology, but very little experience in the cable industry," says Nussrallah. The first thing he did was "hire the best damn cable salesman," who is more commonly known as David Nicholas, vice president of sales for Concurrent's XSTREME unit. From there, Nussrallah's onus has been to build up the division and to start making hay.
Ever since, Concurrent has been gathering industry confidence and racking up deals.
"In 1998, we executed the first agreement with Time Warner to be part of the Pegasus program," says Nussrallah, "and that's what actually kicked it off."
That deal eventually turned into two Time Warner system deployments that serve a combined 1.1 million basic cable customers: Tampa Bay, Fla. and Oceanic Cable in Oahu, Hawaii.
Cox, meanwhile, has selected Concurrent for commercial deployments in San Diego and Phoenix, which eventually will equate to more than 1 million subscribers passed for VOD. Concurrent also will work with Comcast on two of the MSO's three VOD market trials.
Moving forward, Concurrent plans to target big MSOs and big markets. However, smaller, independent cable operators also could be among Concurrent's future customers. "Depending on where they fall in with digital deployments, (small cable operators) would be candidates for our VOD platform. But right now, we're focusing on the big ones, because that's where the big bucks will be down the road," asserts Nussrallah.
To enter larger markets, Concurrent also must ensure that its platform interoperates with myriad classes of digital-set tops and operating systems. Today, Concurrent has integration deals with Motorola, S-A and Pegasus-compliant boxes such as the Pioneer Voyager.
"We've also shown integration with Philips boxes and are in various talks with other set-top vendors such as Sony and Pace," says Del Kunert, vice president of marketing for Concurrent's XSTREME division.
At the same time, Concurrent's platform also must converse with cable operator billing systems. "If you can't talk to the billing system," says Kunert, "then you don't have a business."
Concurrent struck a deal with CSG Systems Inc. to integrate its BackOffice Software with CSG's Broadband Express customer care and billing solution. Concurrent also is in discussions with other leading billing software vendors, including Convergys and DST Innovis.
The work won't stop there. "My challenge is to continue our leadership position," explains Nussrallah. "At the same time, I'll work on other relationships and partnerships to ensure that we're compatible. That includes getting more and better content."
Though the VOD market is forecasted to erupt over the next five years, Concurrent knows its deal clock is ticking even faster.
"Our market share will be largely determined within the next 18 months," says Nussrallah.DIVA
DIVA, which is in the throes of a pending initial public offering that could raise as much as $75 million, blasted from the blocks in 1995 and initiated its first field trial in 1997 with Comcast in Philadelphia, Pa.
During its early days, DIVA collaborated on three other VOD trials: Adelphia in Lansdale, Pa.; Cablevision Systems in Monmouth, N.J.; and Rifkin in Roswell, Ga. Though all four of DIVA's initial trials have since ended, they were all deemed "successful."
Indeed, DIVA's test with Cablevision in Monmouth grew to a lofty 1,500 VOD subscribers. Today, DIVA is in discussions with Cablevision for potential VOD and interactive program guide deployments, says Tim Ray, DIVA's chief operating officer.
Experience culled from those market tests, the company says, has paved the way for aggressive commercial deployments with Chambers Cable (soon to become AT&T Broadband), Charter Communications, Insight, MediaOne (now AT&T Broadband) and U.K. cable operator NTL Inc. DIVA can also expect to get some deployments with Time Warner Cable after the two companies signed a VOD integration agreement for the Pegasus platform.
At the same time, DIVA has had to grapple with how the industry perceives the company and its turnkey approach of providing a full suite of its own hardware, software and back office systems. "Originally, we were thought to be very proprietary with a closed system," says Ray.
That's not true anymore. That same turnkey approach is beginning to bear fruit as DIVA integrates its platform with various set-tops, program guides, middleware and operating systems. For example, DIVA has recently cobbled together deals with OpenTV and Liberate, resulting in integration pacts and a combined $9 million investment in the VOD company. DIVA also is developing a prototype that runs on the Microsoft TV platform. On the billing front, DIVA interoperates with the big three: DST Innovis, CSG and Convergys.
In MediaOne's (now AT&T Broadband's) Atlanta system, meanwhile, DIVA has worked tirelessly to finalize its integration with TV Guide's electronic program guide, setting the table for a VOD ramp up in September or October 2000.
"That's just been a long discussion. Everything but the DIVA—TV Guide part of this has been ready to go for some time," explains Ray. "We're happy to say that we're there, and the easy stuff—completing the integration—is left."
Though that deployment has taken longer than others, DIVA expects the gap to narrow as the MSO launches VOD in more systems. For instance, Insight's second and third VOD launches moved along much more rapidly than the first one did.
"There's a real learning curve here; it's not like adding another pay channel," says Ray. "The timeframe accelerates as we go along."
Instead of relying on a content aggregator like Intertainer, DIVA prefers to go directly to the source. Today, DIVA's content providers include big name and independent movie studios such as Dreamworks, Disney, Universal, Artisan and Unapix; "special interest" providers such as Discovery Channel, CNN and ESPN; children's programming and adult entertainment.
While DIVA has been making the grade in the U.S., its aspirations are on a global level.
In May, NTL Inc., which recently snapped up Cable & Wireless' cable properties, made a $6 million equity investment in DIVA, opening the door for DIVA in the U.K. and the rest of Europe.
"We expect to be putting service in two headends, one with a Cable & Wireless system and another with NTL, and those should launch in July or August," says Chris Good, DIVA's executive vice president and CTO. "NTL is making a big commitment to VOD, and a big one with us, hopefully."
That initial excitement has prompted DIVA to open up a U.K. office, which today houses seven employees. "We're expecting big things in the U.K. and on the European continent," adds Good.
Like its rivals, the forthcoming challenge for DIVA is to keep the VOD ball rolling. "Our second-generation products are moving into the marketplace right now," says Ray. "We have to continue to improve the metrics for cost and performance. That's our challenge, as it is for everyone."DemandVideo
DemandVideo has taken a different tack than most of the VOD players out there. Instead of building and developing its own equipment, it buys servers and then markets a turnkey VOD platform to overbuilders and incumbent cable operators.
"We buy all of the equipment," says Susan Marr, DemandVideo's vice president of marketing. "The operator doesn't have to pay for anything, except maybe some network elements like QAM modulators, depending on what server we put in there and what set-top box or platform they're using."
Instead of relying on money generated from VOD equipment sales, DemandVideo gives the movie studios a piece of the revenue, and splits the remaining gross revenue with the service provider.
"We're attractive to operators that don't want their own equipment, and want a turnkey service," says Marr. "Some of our clients may want to buy their own servers, and we'll be flexible with that, too. It would just change the revenue model a bit."
Currently, DemandVideo's only deployments are with Seren Communications' overbuild systems that serve customers in St. Cloud, Minn. and Contra Costa County, Calif. DemandVideo also has a pay-per-view deal with Knology, an arrangement that eventually could migrate to VOD.
"Within the next year," says Marr, "we'll be launching in about 45 market segments. Those markets include Latin America, as well."
Though overbuilders currently fit the bill for DemandVideo's VOD endeavors, the company also will seek deals with incumbent MSOs if their systems meet two qualifications: a two-way digital plant and a marketable bundle of voice, video and data services.
"If an MSO meets those criteria, we would certainly work with them," says Marr. "What has excluded us in the MSO market is that most of them are making the decision to do (VOD) themselves, although I expect that some will want to do a make-versus-buy decision for VOD. Right now, new network builders have been a priority for us, because they have two-way digital systems."
Those digital systems will be outfitted with nCUBE servers, at least for now. DemandVideo cut a deal last year to use nCUBE's equipment as the company rolls out VOD. That agreement is non-exclusive, however.
"I could see us buying Concurrent and other servers," says Marr. "So far, what's been preferred is nCUBE, because our integration with them has not been a barrier."
While DemandVideo has its source of equipment solved, getting enough content could become a challenge as VOD reaches higher penetrations.
"The thing that's on the top of my list is grappling with not getting as much content as we would like to have," explains Marr. "Everyone has the same problem, so packaging and marketing will be the challenge to get the buy rates we want."
In the meantime, DemandVideo plans to bolster its VOD offering with local content, reserving some space for operators to offer high school football games or even city council meetings on-demand.
"We would just charge a minimal fee for storage and encoding. A lot of operators we work with are community-centric, so something like this would give them a local flair to their VOD product."nCUBE
What started out as a 1983 development effort to produce massively parallel processing computer systems for the scientific and engineering markets has, according to nCUBE officials, been transformed into a massively powerful way to bring Arnold, Harrison and Mel into homes anywhere in the world at anytime with a click of a button. Jay Schiller, nCUBE's director of product management broadband solutions, says the company made a crucial change of course about seven years ago when it veered away from scientific computers and zeroed in on video streaming.
"There are things you can do with a general purpose computer for pumping video," says Schiller, "but you have limitations. If you focus on the system, not as a video server that's based on a file server pumping video, but as a streaming media system, i.e., an application-specific appliance for pumping video, you can build extremely reliable, highly dense and scalable video server solutions."
nCUBE's VOD solution is based on hypercube architecture, which it had developed for the scientific and engineering communities, and has been incorporated into what it calls its Streaming Media Appliance. Its newest "appliance," the MediaCUBE 4, is the fourth such video server aimed at cracking the VOD nut. Schiller says the MediaCUBE 4 has been designed to offer VOD service providers with the flexibility, scalability and functionality they'll need as VOD ramps up and rolls out.
To accomplish those lofty goals, says Schiller, the MediaCUBE 4 holds 12 hot-swappable ,9, 18, 36 or 73 GB drives. In the back, it holds up to five PCI cards for a variety of output formats (i.e., QAM 64/256, DVB-ASI, ATM OC-3/12, Fast and Gigabit Ethernet), as well as hot-swappable, redundant fans and power supplies. The MediaCUBE 4 also provides hypercube interconnect ports that allow providers to connect additional servers for incremental scalability on a massive scale.
Schiller says one media hub will stream 500 Mbps. With the interconnect ports, providers can link up to 256 of the servers, functioning as a single computer, streaming up to 128 Gbps. How does that translate into VOD? Schiller says that with one piece of content (e.g., one of those Arnold, Harrison or Mel flicks), in an MPEG-2 stream being delivered at 3 Mbps, "you could be pumping 44,000 simultaneous streams, unique to different subscribers, from the same (single) piece of content."
This daisy-chain capability, says Schiller, also provides distribution flexibility. "We could be deployed in a very large system centrally in the headend serving many optical nodes and subscribers," explains Schiller. "Or, we could be located in a distributed manner, in the hubs. Or we can even be a combination of both. For example, you could have more popular titles in the smaller systems down at the hub; less popular titles in a centralized system at a more convenient location. Or, you could be streaming from one hub to another."
While there may not be any VOD providers hooking up 256 MediaCUBE 4 servers just yet, Schiller says the stage is set for a multi-faceted VOD battle. As far as the cable operators go, says Schiller, "The priorities for VOD are being set by the competition from DBS, the telcos and other overbuilders. At the same time, on the telco side, they see VOD as a key service to compete with cable, overbuilders and DBS. Over the next 12 to 18 months, it's going to be a nuts-and-bolts rollout period. You're going to get past trials and the dipping of toes in the VOD waters. The trial period is ramping down now, and the install period is going to ramp up rapidly."
Even the trial period, says Schiller, is something of a misnomer. VOD as a fait accompli, he says, has definitely settled in. "Everyone seems to realize this is a reality. The question now is more about when are they going to do it. I think the general perception is that it's technically feasible. In fact, the trials are not really technical trials. They're typically marketing trials. How am I going to get into this business? What do I need to do? How do I need to do it? It's not, 'Does your technology work'?"SeaChange International Inc.
Given the nature of the cable beast, as it were, Yvette Gordon, vice president of interactive technologies at SeaChange International Inc., is not all that surprised that industry engineers are fascinated with the hardware in VOD systems. But like so many other technological developments today, VOD's magic, says Gordon, isn't so much in the boards and silica of the equipment, but in the software applications that make it work.
"It seems a lot of people seem to focus on the hardware—there's a video server and a set-top box, for example," says Gordon. "But the software aspect of how content gets around and how it goes on and off, and how easy it is to manage, and how easy is it to troubleshoot, is very important."
She notes that people are understandably focused on the capital purchase and ROI of VOD systems as well. But here again, the cold, hard cash aspects of the business, much like the cold metal of the hardware, is just on the surface. There are other subtle, but no less important, aspects to consider.
"Part of that ROI, in my opinion, is really the operational cost of just how many people is it going to take to operate it," says Gordon. "How much of an impact is it going to have? And then, there's the more subtle aspects to consider.
"For example, if a customer has a problem, how difficult is it for that problem to be overcome? Does the system have the ability to solve it itself? If someone has to get involved, are they automatically notified? Is it easy to fix the problem? These are things that can't be readily assigned a monetary value. It's part reputation. It's part operational ease and making this system something that a cable technician can really manage."
When it comes to VOD hardware and software, SeaChange's ITV System is an ingenious blend of both, says Gordon. The ITV System is based around the company's MediaServer technology, which is comprised of multiple computers directly linked in a high-speed mesh architecture that effectively serves as a single server. Single copy storage is achieved through the company's RAID2 and MediaCluster technology that uses data striping and parity techniques across disks in a computer, as well as across the multiple computers in a server.
The ability of a VOD system to utilize just one copy of content to feed multiple requests at different times begs the question of peak loads and how robust a system should be to address such loads. Conventional wisdom says VOD systems should be designed, at a minimum, to handle 10 percent of the subscription base at peak. Gordon, who had a hand in developing that wisdom as an engineering director at the famed Full Service Network (FSN) trial in Orlando, Fla., says that number is not set in silica.
"The 10 percent number, which was made public in 1997, came from the Full Service Network that Time Warner built. The FSN was built to 25 percent and the recommendation at the completion of the trial was that 10 percent was an optimal number.
"I really believe that number is a good number for offering a variety of services. I think, however, if you're going to run movies only, that 10 percent number could be high. However, when you get to really high penetration rates of digital set-top boxes, and you have potentially true all-broadcast, time-shifted programming across a very wide variety of channels, then I think that 10 percent may be too conservative."
The apparent prevailing notion among operators and vendors alike that basic VOD technology is essentially locked in, and it's only a matter of deployment, says Gordon, is a little simplistic for a very complicated service that's barely gotten out of the headend. She notes, for example, that there needs to be a standard for encoding VOD content so that content aggregation can be simplified and done faster.
"I think where the industry is going with video-on-demand," says Gordon, "is that there has to be a standard for encoding so that a post-production house can directly encode its own content for VOD. Potentially, the licensing would be like it is with any other type of content. That way, you could get an on-demand DVD or VOD movie as one of the standard types of content directly from a studio. For the industry, it would make on-demand content aggregation much easier."
Gordon, like her VOD peers, realizes her company cannot stand still. The broadband industry is moving too fast for that. While VOD technology may have matured a great deal these past 18 months, the challenges for on-demand service providers like SeaChange, nCUBE, DIVA and Concurrent have not lessened. In fact, says Gordon, there's plenty of them to go around for everybody.
"Now," says Gordon, "the technical hurdles are about taking the next steps. It's about the services that are going to come tomorrow.
"What does that include? It's about real-time billing interfaces for e-commerce. It's about mixing Internet applications with VOD. It's looking at imbedded URLs in VOD streams and skipping back and forth between broadcast programming, VOD and interactive television. The whole thought of PVR (personal video recorder) capability from video servers is another issue. On top of all that, there's always the issue of cost reduction. How do you make VOD more cost-effective as it reaches large-scale rollout status?"
Obviously, while VOD has come of age a lot faster than many had expected, there's still quite a way to go before it settles down.