What’s an Edge QAM, Again?
by Leslie Ellis // November 27 2007
The quintessentially techie “QAM,” for quadrature amplitude modulation, is back in the engineering limelight, this time with a new prefix: Edge.
When you hear people talking “edge QAMs,” chances are high that you’ll hear the term “switched digital video” within a few sentences. That’s because digital video switches created the need for edge QAMs.
It goes like this: All digital cable services carried over cable plant use QAM modulation. It’s the conveyor belt. It’s what moves video, voice and data services from headends to homes.
(Brush-up basics: People tend to pronounce it as a word — “kwahm,” as in, rhymes with “Guam.” As for physical location, it’s a headend/hub thing: An unassuming, rack-mounted metal box, enclosing a series of slide-in cards. Carrying capacity goes like this: One QAM equals 38.8 Mbps of downstream data, or two to three HD streams, or 10 to 12 regular digital video streams. It is the equivalent of one analog channel.)
Up until recently, cable providers purchased QAMs on a service-specific basis. The QAM assigned to move that Web page into your cable modem, for instance, couldn’t also be used to move those digital video channels into your high definition TV.
Likewise, a QAM dedicated to video on demand (VOD) couldn’t also be used for switched digital video, even though their architectural constructs are identical.
Enter the edge QAM. It’s built to carry both VOD and switched digital video streams. Ultimately, the edge QAM will move three categories of service (those two plus the Internet Protocol side of the house, meaning data and voice.)
When people say “edge QAM,” then, they mean multi-purpose. They aren’t talking about where it is, physically — at the “edge” of the network, wherever that is. (For directions, see “A Pocket Map to the Edge of the Network,” in the 03/15/05 edition.)
In the big picture of cable technology development, the edge QAM is fairly unobserved — but bubbling with supplier activity. By my count, seven companies are building the gear, up from two a few years ago. Among the providers: Arris, Big Band Networks, Harmonic Inc., Motorola, RGB Spectrum, Scientific-Atlanta/Cisco, and at least one skunk-works outfit.
Cable operators look for three things when evaluating edge QAM innovation: Price, density, and, for lack of a better term, “open-ness.”
Price is predictable. Right now, QAMs run in the $250 range. The goal, however ambitious, is to lop off the zero.
One rack-mounted unit typically contains cards that operate between eight and 24 QAMs. That’s the “density.” One pointed area of innovation, then, is upping the density. Higher density, lower price.
And then there’s the “open-ness” piece. Switching is a big move. All major cable operators are or will be doing it as a way to preserve bandwidth. None of them want to get painted into a corner with a single, monolithic supplier, seeking to control the economics of what happens beyond the switch. They’ve seen that movie before, and they don’t like the ending.
An edge QAM is considered a network resource. It talks to at least two other devices: The “session manager,” which sets up the linkage between you and the VOD server (or the switch), when you choose to watch something, and the “resource manager,” which figures out which QAMs are supposed to be moving what stuff to where.
An industrial movement is well underway to open up the conversations (the “protocols”) between those linkages, so that operators can buy things modularly. A switch from one guy, a session manager from another, a resource manager from a third, and edge QAMs from whoever is the most dense, open, and affordable.
Mix and match is the name of the game.
From an overall cost perspective, edge QAMs matter because they currently represent something like 70% of the capital spend for deploying switched digital video, according to the MSO-side technologists who track such matters.
By opening things up and making them more modular, operators reason, the actual QAM costs come down, which brings that percentage down, which makes switching more affordable, which make Wall Street happy. Lather, rinse, repeat.
This column originally appeared in the Platforms section of Multichannel News.
The Bandwidth Trifecta for Program Networks
by Leslie Ellis // November 12 2007
If your company moves high definition TV to its distributors by way of satellite, there’s yet another bandwidth saving superstar to help you get 10 more pounds of HD into that five-pound bag.
What it is: A new way of saving room on those spendy satellite transponders, which can cost a content owner upwards of $125,000/month.
How big of a bandwidth savings? About 30%, say tech people at program networks and aggregators. One content-side engineer explained it this way: “Our transponders right now can carry 47 Mbps of video, and the S2s will carry somewhere between 65 and 72 Mbps.”
That extra 18 to 25 Mbps makes room for another two or three HDTV streams per transponder.
In other words, the satellite transponder that currently carries two or three HD streams (using existing MPEG-2 compression and existing modulation/coding techniques) could now carry five or six (using advanced compression and new modulation/coding techniques).
“The “S2” in that mention, by the way, is verbal shorthand for “DVB-S2,” which is the bandwidth saving superstar of this week’s translation.
“DVB-S2″is the name of a technical standard. It breaks down like this: “DVB” stands for “Digital Video Broadcasting,” and is the major standards-setting body in Europe. “S” is for satellite. The “2” is the version number of the standard.
(Version one of the DVB-S standard, as a point of reference, blueprinted more than 100 million digital satellite receivers, worldwide. That’s since 1993.)
The underlying context of DVB-S (versions 1 and 2) is all about how to anticipate (and thus prevent) the errors that inevitably occur when blasting bits up and down from space.
From here, the DVB-S2 detail can corkscrew way into the tech-funk. (This is generally true of anything involving satellite technologies. If you find cable tech-talk daunting, spend an afternoon with a satellite engineer.)
Here’s an example from an email last week: “The LDPC codes replace the Viterbi Forward Error Correction (FEC) of DVB-S, whilst the Reed-Solomon code is replaced with a different BCH (Bose-Chaudhuri-Hocquenghem).”
How DVB-S2 Works
How it works requires a brief refresher on how satellite transmission works. For starters, the modulation has to be sturdier, because sending stuff up into space is much harsher than sending stuff over a wire. Satellites traditionally use a form of modulation called “QPSK,” for “quadrature phase shift key.”
Don’t get hung up on the language. To put it in context, the QPSK modulation used for space transmissions is also used by cable systems to transmit in the upstream (home to headend) direction. The upstream path of cable is theoretically as hostile (in different ways) as space.
When you’re sending stuff into a hostile environment, even if you’ve taken extra precautions when imprinting it onto the carrier (modulation), you need to carve out room for reconstructive surgery — just in case something gets really mutilated during the ride.
The safety mechanisms are known as “forward error correction,” or FEC. Techniques vary. Understanding them involves heavy math. Short version: Send extra bits that know where they’re supposed to go, if they get called upon on the ground to stand in for missing data.
DVB-S2 is an improvement in forward error correction, which harnesses improvements in modulation. It’s already in use by some content owners and aggregators. It’s not something that can be easily adopted by DirecTV and EchoStar, for legacy reasons: On the receive end, it requires gear that can demodulate and de-code in the new way.
Conversationally, DVB-S2 tends to move in step with MPEG-4 compression. The reasoning: You’ll need new gear to do MPEG-4, and you’ll need new gear to do DVB-S2. Might as well make both changes at once.
And so ends this unintended mini-series on HDTV and bandwidth. To sum it all up: People who pay to move big HD streams over satellite will soon get a trifecta of bandwidth savers.
One is advanced compression, like MPEG-4. Two is better modulation, to move stuff up and down from space. Three is better error correction, to reconstruct streams on the ground.
This column originally ran in the Technology section of Multichannel News.