| What Does It Take To Launch DOCSIS 3.0? |
<Translation Please 2.0>
Many moons ago, when cable providers discussed what new technologies they'd roll out this year, one in particular ranked high on the list: The newest chapter in the cable modem chronicle, known as DOCSIS 3.0. (Note: Some people now shorten the term to "D3.")
And here we are, with three months left in the home-run stretch to year-end 2008.
This is probably why the email basket grows more full each week with questions like this: "What are you hearing about what it will take for cable operators to take DOCSIS 3.0 from trial status into a production network?"
If you haven't gone there yet on your list of things to understand better, know going in that the assortment of answers to this question is bountiful.
For that reason, this week's translation aims to begin to untangle the DOCSIS 3.0 launch sequence.
Just Swap a Few Boards
The glamour answer to the question of what it takes to launch DOCSIS 3.0 sounds deliciously effortless: "You just swap a few boards in the CMTS, or populate the empty blades, depending on whose gear you're using."
Alas. That word "just" can mislead. Best to go deeper.
Here's a sampling of the range of answers and observations from my notes, about what it takes to launch DOCSIS 3.0.
The best-practices answer, from engineers already active in DOCSIS 3.0 trials: Don't skip versions. Going directly to DOCSIS 3.0 from DOCSIS 1.x, within the same scheduled maintenance window, will register high on the scale of winces.
The bandwidth answer is to be mindful of available shelf space. If your main intention in going to DOCSIS 3.0 is to bond two or more 6 MHz channels into a faster, bigger pipe, remember that those channels need to first be empty -- or at least not used for anything but broadband traffic.
[Part of the bonding also happens at the other end of the network, inside the new DOCSIS 3.0 modems. The number of channels bonded is a function of how many tuners are used inside each modem -- two tuners yield a two-channel bond, three tuners, a three-channel bond, and so on.]
And then there's the labor answer. Here's a verbatim example from a recent batch of notes: "The bulk of the work [in going to DOCSIS 3.0] is the re-lash -- there's a fairly significant amount of recombining."
Let's unpack that. To "recombine," in this context, is to take a closer, more strategic look at how certain frequencies are smooshed (combined) together, out to which groupings of neighborhood nodes.
Those certain frequencies that get combined to squirt out to hubs and nodes, come from -- guess? -- those ubiquitous digital modulators, known as "QAMs."
And yes. We are about to veer into the realm of network topology. Take a deep breath.
The One-Size-Fits-All QAM
In the language of QAMs, node groupings are usually called "Serving Groups" or "Serving Area Groups." Rule of thumb: Four 500-home nodes equals one serving area group (see the April 10, 2006 and April 16, 2007 translations for more detail).
In the beginning, every digital cable service came with its own blend of QAM modulator: Digital video and VOD had one kind. High-speed data had another, and so on. They all did essentially the same thing -- imprinted digital information onto carriers to be transmitted to homes. But different services attracted different vendor groups for the QAMs.
Likewise, because digital video, VOD and high-speed data services launched in different timeframes, the ways in which their respective frequencies (QAMs) were combined and sent to nodes wasn't parallel.
Then, somewhere along the way, cable operators learned that they were paying way more for data QAMs than they were for video QAMs. (Way.)
Why, MSOs asked their suppliers, can't we buy QAMs that can do video, voice, or data, all in the same unit? Is there a one-size-fits-all QAM in my future?
Work commenced to find ways to whack those extra zeros off the price tags of data QAMs.
Now that operators are starting to think through the launch sequence for DOCSIS 3, these combining conversations are again in vogue.
Operators active with switched digital video (SDV), for instance, combine the QAMs feeding into the switch in the same pattern they combined QAMs for VOD. Meaning they made the combiner groups for VOD and SDV supply the same groupings of four 500-home nodes, or 2,000+ set-top boxes.
Likewise, operators readying for DOCSIS 3.0, and who want to steer toward "universal edge QAMs," may also want to group the frequencies (QAMs) of the bondable channels in parallel with VOD and SDV.
In other words, operators may want to "re-lash" the DOCSIS QAMs so that they, too, feed the same groupings of four nodes, or 2,000+ cable modem households.
Re-lashing is a wiring thing, which means it's a labor thing.
The To-Do for DOCSIS 3.0 -- So Far
So far, at this point in the transition to DOCSIS 3.0, the to-do list is as follows:
Make sure you have the boards ("blades") you need for the CMTS. This varies by vendor.
Make sure you can get 3.0-based modems.
Go to DOCSIS 2.0 before you go to DOCSIS 3.0.
Before you get out the glue, so to speak, for channel bonding, make sure you have the channels to bond.
And, if you're on a path toward one-size-fits-all QAMs, look now at whether you want to align your wiring to be consistent across all digital services -- linear and on-demand video, switched video, and broadband / IP services.
This column originally appeared in the Platforms section of Multichannel News.
