Technology

Translation Please: The Problem of Units, Part Two

8/02/2009 2:00 AM Eastern

Reader Charlie, a veteran cable engineer, chimed in last week with another angle to the problem of units.

It is this: When it comes to broadband, if you think it’s hard to get your head around per-household usage, try conjuring per-system throughput.

“I’m still not sure the world will ever get a handle on knowing how many bits and bytes they use,” he wrote. “I break my calculator when I try to extrapolate this to how many bits we deliver every month for $99.”

Charlie’s math goes like this: The cable operator who maintains 135 digital channels, using 256-QAM modulation, is actually sending 5,670,000,000 bits of information every second — “most of which falls on the floor, unused.”

Let’s unpack his logic. Why 135 channels? Because from a bandwidth perspective, that’s the whole shelf for the operator built out to 860 Megahertz. The math goes like this: Subtract the bandwidth of the upstream path (54 MHz) from the top (860 MHz); divide that answer by six (the measure of a traditional “channel” size, in MHz).

What’s that 256-QAM doing in there? In this case, it’s a measure of carrying capacity, and means that each of those 135 digital channels has the throughput equivalent of around 40 Megabits per second, or Mbps.

What he means by the bits falling on the floor is that people generally watch TV one channel at a time. Thus lots of channels go unwatched, most of the time.

Charlie’s right, of course. The numbers around broadband supply are just as arcane as the numbers defining broadband usage.

The blessing and curse is this: The total carrying capacity of any cable system built to 860 MHz is 5.6 Gigabits per second (Gbps), reusable across nodes. Important: This measure only works if all 135 channel slots are empty.

Let’s go crazy and say they were. Given today’s broadband consumption patterns, that 5.6 Gbps should keep everybody happy for at least another two weeks. (Kidding, but still …)

Stumped by gibberish? Visit Leslie Ellis atwww.translation-please.com.