It’s a dense, dense, dense, dense world for universal edge QAMs
Universal edge QAMs are taking on added importance as cable deploys
more SDV, VOD – including Start Over, Look Back, Catch Up
and nDVR – and DOCSIS 3.0 services.
Driven by the convergence of various forces, universal edge QAMs are poised to step up onto the next rung of the evolutionary ladder.
While universal edge QAMs have been around for a relatively short time in cable years, they’ve taken on added importance as cable deploys more switched digital video, VOD – including Start Over, Look Back, Catch Up and network DVR – and DOCSIS 3.0 services.
Throw in the fact that most cable operators are moving toward all-digital, which means even more on-demand content with the freed-up bandwidth, and there’s a clear need for even denser edge QAMs. Cable operators not only want denser edge QAMs per port, but also lower power consumption and less rack space.
“We're definitely looking for these products to get more dense,” said John Civiletto, Cox’s executive director of core video systems. Referring to the newer generation of QAMs, he said: “These definitely represent a revolution in the edge QAM. Devices in the past weren’t comparatively very dense. These offer maybe a two- to four- to eight-times increase in density, and we expect over the next five years or so that we’ll see a requirement in the edge of our network growth by a similar factor.
“Part of the attraction of very dense edge QAMs is to put multiple services on the same edge QAM. So you want to be able to put your DOCSIS, your VOD and SDV traffic all on one edge QAM, and that is only now starting to become a reality from an application point of view. It’s definitely a case of we’re right on the edge of what is practical in addition to where the products are.”
Charles Cheevers, Arris’ vice president of multimedia product management and European CTO, said SDV can take up four to 16 QAM channels, VOD four to eight channels, and DOCSIS 3.0 four to eight channels, but that those numbers are changing.
“If you look at the breakdown of unicast services (SDV and VOD), it would be low, somewhere at 16 channels per service group at the moment, up to 32 on the high end as more unicast services are required for more personalized TV content,” he said. “As more and more DOCSIS 3.0 gets rolled out, it will go to eight, and then possibly 13. You now start to need a lot more QAM channels, so we’re trying to get that cost per downstream as efficient as possible in order to prepare for that evolution of QAM channel requirement.
“The bottom line is if you add all of them up, you get a very large number. If you don’t add them all up and have them separated and siloed, which is typically the case at the moment, you don’t get the cost-effectiveness of putting them on one F-connector.”
Doug Jones, BigBand Networks’ chief architect, said the increasing number of QAMs needs to be denser in order to fit them in the existing distribution hubs, some of which date back to the mid ’90s.
“There’s been an explosive growth in QAMs, from 20 per distribution hub back when everything was broadcast to 1,600 per distribution hub for narrowcast on-demand, switched digital video and DOCSIS services,” he said. “The distribution hubs aren’t going to get any bigger because no one wants a bigger distribution hub in their neighborhood. The QAMs are a clear candidate to get shrunk down and more dense, and the industry has been doing that.”
FEELING A LITTLE DENSE?
While there’s a clear need for denser edge QAMs, there are some differences of opinion among vendors as to how dense they need to be.
“I don’t think ‘dense edge QAM’ really does it when you’re talking four to eight QAMs to a port being dense,” said Tony Pierson, LiquidxStream’s vice president of business development. “We use ‘ultra dense’ because there has to be some differentiation when you get above 36, or 70, or whatever the numbers are going to be in the future.”
By using its own silicon, LiquidxStream has gotten out of the starting blocks with an edge QAM that supports 36 QAMs per port, for a total of 576 QAMs per chassis, and Pierson said 70 QAMs per port was reachable. With its density, LiquidxStream is able to devote an RF port to each service group for SDV, VOD and data.
“We have 36 right now because that made the most economic sense in terms of the amount of capacity that you need on the box and the amount of growth you want to allow for,” he said.
Pierson said LiquidxStream intitially programmed its chip on a field-programmable gate array (FPGA) before switching to an application-specific integrated circuit (ASIC). By using an ASIC, LiquidxStream was able to reduce the power of its LxS-3616 Edge QAM, which in turn cuts down on the demand for air conditioning and lowers the cost of HVAC, backup power and generator capacity.
Pierson said other benefits to using his company’s ultra dense QAM include lowering the number of coax and optical Ethernet connections.
“Once we got out into the field, we were able to show how much you save by going to a higher density,” Pierson said. “You’re sharing the RF output technology, which sucks up a lot of power, over more QAMs, so you can divide that up by a bigger number and get a lower number per QAM on power consumption.
“You don’t need as many optical connections because you’re connecting to fewer boxes and not as many coax connections because each service group has its own RF port. When you need to upgrade, you don’t need to roll trucks to neighborhoods or spend a week upgrading from two QAMs to four QAMs. You do it sitting at a desk, and that saves you money.”
Pierson said LiquidxStream was in all of the major labs for testing, but he couldn’t name the cable operators that were taking the company’s edge QAM for a spin.
Not surprisingly, vendors are cruising along different tacks when it comes to improving on their universal edge QAMs. Earlier this year, Arris rolled out a new high-density QAM-RF module for its D5 Universal Edge QAM. The Quad Port Module 8DX4 has four RF ports, with each able to output up to eight QAM channels.
“We’re focusing on four to eight channels as the sweet spot for the products we’re shipping this year,” Arris’ Cheevers said. “With eight channels per F-connecter, we fit into almost every single application in a cost-effective manner, and pricing in the QAM business has gone down because of that ability to deliver eight channels.”
“If we were having this conversation 18 months or two years from now, I’d say 32 channels is probably perfect, but in ’09 it would be real hard to find an operator anywhere in the world that would deploy an F-connector that had 16 channels for combined services or applications,” added Derek Elder, Arris’ senior vice president of product management and marketing.
In June, BroadLogic said its TeraQAM BL85000 chip for QAMs would be available for purchase later this year. The TeraQAM will be available in 16-channel and 32-channel versions, and like LiquidxStream, it uses an ASIC instead of an FPGA, and then uses a less-expensive FPGA for other system functionality.
BroadLogic, whose investors include Time Warner, Comcast and Cisco, also touted the use of its chip in helping cable operators transition their networks to IP in order to maximize the investment in their fiber deployments (see Figure 1).
FIGHT THE POWER
With less space in distribution hubs, edge QAMs not only need to be smaller and denser, they also need to be more energy-efficient, the latter of which is top of mind for both QAM vendors and cable operators.
“We’re definitely putting an emphasis on reduced footprint and reduced power consumption,” Cox’s Civiletto said. “Power consumption is a big opportunity. It has some green benefits that Cox as a company is interested in pursuing, and it also helps us in managing our facilities.”
Cheevers said Arris’ QPM 8DX4 does 2.7 watts per channel, which is down 5 to 6 watts from previous generations.
“Obviously that’s a function of how dense it is, but it’s also a function of how you code your FPGAs,” he said.
In the same vein, Pierson cited a recent case study by LiquidxStream that found its edge QAM consumed 25 percent less power than the legacy units in a hub.
“We’ve done studies that show after you invest in the original technology base, you can save enough to almost pay for the incremental edge QAMs because operational and power savings are that significant,” Pierson said. “I think there is a lot of value that customers can get from going with a higher density.”
There are other benefits to the next generation of universal edge QAMs, including reduced inventory since the QAMs can be shared between data, VOD and SDV services.
Another benefit of dense edge QAMs is that the cost is going down. Last year, the price per QAM was $400; this year it’s around $200, and it could eventually hit the $100 mark.
“We’re usually talking about a multiplier of three, or even five in terms of taking all of the QAMs that have been deployed over the last eight years that will be deployed over the next three years,” said Gil Katz, Harmonic’s senior director of cable solutions. “They don’t necessarily want to spend a lot more out of their budgets to be able to deploy so many QAMs, so the density of those QAMs and everything related to the operator needs to shrink to make the case."
BREAKING DOWN THE SILOS
Universal edge QAMs also hold the promise of being able to share a QAM dynamically between VOD, SDV and data, but it may be a few more years before that happens. Harmonic’s Katz said the video and data service groups need to be aligned, and there currently isn’t a universal resource manager that can allocate bandwidth for both data and video services.
Another obstacle is that the QAM interleaver for data delivery is 32/4, while video’s interleaver is 128/4, the latter of which Katz said won’t work for data. As for 32/4, it may not work with some set-top boxes.
“Even before service groups become aligned, there are still benefits from the dense edge QAM products that are starting to hit the market right now,” Cox’s Civiletto said. “Even though you can’t directly share between video and data services, it gives us the ability to have a single device that serves both products. That’s one device in your headend instead of multiple devices – that’s less wiring and less power consumption. There are still benefits prior to reaching the end game.”
And there are additional evolutionary steps for edge QAMs to scale. Arris’ Cheevers said direct digital synthesis (DDS) for direct RF reflects the cutting edge of technology since there is no upconversion.
“With direct RF, you go straight out from a digital-to-analog converter, and all you have to worry about is power amplifying the signal,” he said.
BigBand, which is coming out with a new edge QAM early next year, is looking at high-speed digital-to-analog converters (DACs) from companies such as Maxim and Analog Devices.
“They are becoming available, and they can be used in QAMs,” Jones said, referring to the new generation of DACs. “They dramatically lower the cost per QAM and the power per QAM by a factor of two to four, depending on how good your hardware team is. DDS truly is a disruptive technology with what it can do, so I would expect some new entrants to come in. You’re going to have competition going on with the usual players, plus the new entrants using this technology, so there will be a lot of competition out there for this business.”
As far as pricing goes, Cox’s Civiletto said he expects to see sub-$100 QAMs; and with Moore’s Law in place for QAM deployments and increased competition, he may see that price soon.