NCTA Tech Papers, Part 2
by Leslie Ellis // June 11 2012
Last week we took the short route through the longest-ever NCTA Tech Paper, a whale of a thing, at 182 pages.
This week, the bigger trends arising from this year’s engineering scribes, starting with the colorful analogies: “The prisoner’s dilemma” and “the ships passing in the night problem,” used in different papers to illustrate that adaptive bit rate streaming clients inside PCs, tablets, and other screens playing IP video usually work in isolation of each other.
So that multi-megabit-per-second pipe into the house might pour a fat stream to the smallest screen, and a skinny stream to the HDTV – because the two screens don’t know of each other’s existence, under the same roof.
It’s a “prisoner’s dilemma” (handily abbreviated “PDIL” by Arris authors Carol Ansley, Jim Allen and Tom Cloonan) because “clients are faced with electing to optimize (bandwidth) for their own benefit, or they can optimize for the common good of all clients on the network, including their own.”
In a prison sense, then, eat all the food or share it; either way, face the consequences.
The “ships passing in the night problem,” described by Ericsson’s Michael Adams in a lunch meeting prior to the show, and co-written with Chris Phillips, references a classic networking problem, akin to Schrodinger’s cat for mathematicians. It goes like this: Networks work in seven independent layers, known as the “OSI stack.” (People tend to identify themselves by which “stack” their work represents – “I’m a layer 3 guy,” and so on.)
So, by design, what’s transpiring on, say, layer 2, cannot fathom what’s happening up on layer three, or four, and so on, up through layer 7. Like ships passing in darkness.
Weirdest pronounceable acronym: “CAPWAP,” for “Control and Provisioning of Wireless Access Points,” about MSOs and WiFi backhaul. (Authors: Cisco’s Rajiv Asati, Rajesh Pazhyannur, and Sangeeta Ramakrishnan.)
If you saw the “surfaces” exhibit at the show, and wondered how to turn that man-cave wall into a giant TV screen without projectors, check out “Surfaces: A New Way of Looking at TV,” by Kevin Murray and James Walker.
And, go figure, MPEG-4/AVC/H.264 digital video compression turns 10 this year (seems like yesterday). Its successor: “H.265” and/or “HEVC,” for High Efficiency Video Coding.” Just as MPEG-4 did for MPEG-2, H.265 offers 50% coding efficiencies over H.264, explain Motorola’s Robert Howald and Sean McCarthy in “Bits, Big Screens, and Biology.”
What this means, practically: An HDTV stream compressed to 16 Mbps, using MPEG-2, squishes down to 4 Mbps, using H.265. Which seems a good thing for the commercial plausibility of 4K video. The efficiency gains come from “context-adaptive binary arithmetic entropy coding,” or “CABAC.” (Uh-huh.)
This column originally appeared in the Platforms section of Multichannel News.