Cable-Tec Expo: What's the Magic IPTV Number?
According to John Chapman, a Cisco Systems Inc. (Nasdaq: CSCO) fellow and CTO of the company's access and transport technology group, that number will likely fluctuate as cable looks to replicate more and more of its linear programming in the IP domain.
Cable, he said here during the show's opening general session, could start out by bonding just four channels during the early phases of the migration, then move to eight as operators add in a mix of standard- and hi-def channels. A full offering, with 50 percent of the homes passed getting it, might require 16 bonded channels.
Given that each 6MHz channel in use today by North American MSOs supports about 40Mbit/s, that translates to capacity of 160 Mbit/s for four channels, 320 Mbit/s for eight, and 640 Mbit/s for sixteen. The latest and greatest Docsis 3.0 modems can bond just eight downstream channels, so that ground is already covered. D3 modem chipmakers are just starting to dabble with 16-channel bonded downstream capabilities. (See Broadcom Dips Two Chips in 800M 'Prototype' .)
Cox Communications Inc. SVP of technology Jay Rolls agreed that eight channels should be enough to get the IP simulcast going, but said that configuration "might get a bit limiting" if operators couple the IP simulcast with a high-speed Internet service. To account for that, cable will need 16 channels, he said.
The discussion is important because cable will someday migrate its managed video services to IP and perhaps do a better job embracing video that's coming in over the top.
Tied into that is the industry's broader TV Everywhere efforts, which will need to step up as over-the-top threats like Netflix Inc. (Nasdaq: NFLX) establish beachheads on a multitude of IP-connected TVs, tablets, and smartphones. That should give cable an opportunity to develop clients that run on those devices, too.
No one on the panel was willing to say when an MSO's logo and IP-based video service option will start appearing on retail devices like the Playstation 3 and broadband-capable TVs, but "we're already on that trajectory," said Steve Reynolds, SVP of consumer premises equipment and home networking at Comcast Corp. (Nasdaq: CMCSA, CMCSK), which is starting off TVE by piping some authenticated on-demand content via broadband to PCs. "We're already pushing video out to those IP endpoints."
But Rolls suggested that cable should take a measured approach with its support and target CE devices that make the most sense. "Writing a client for every device out there would be futile," he said. "We're not going to be writing clients for every type of device out there."
Another point of debate was when cable might need to widen its upstream path using mid-splits, top-splits, and other techniques that turn exiting downstream spectrum to upstream or look to carve out fresh bandwidth above 1GHz. (See CableLabs Eyes a Super-Sized Upstream and Javelin Jabs at Gigabit Cable Upstream .)
Rolls stressed that cable may not have to establish more upstream for another five years of so while MSOs just get started with bonding upstream channels for Docsis 3.0.
Society of Cable Telecommunications Engineers (SCTE) SVP and CTO Daniel Howard suggested that cable may have to adjust that thinking if it's successful with typically more symmetrical business services that use Docsis.
Time Warner Cable Inc. (NYSE: TWC) EVP of architecture, development, and engineering Mike Hayashi, meanwhile, is not particularly enthused about any upstream moves that could impact service uptime, stressing that "a module swap is not just a module swap." Hayashi didn't offer any specific product examples, but Javelin Innovations Inc. , for example, has introduced a 1.8GHz module/faceplate that can fit into existing taps without plant resizing.
And there are other factors to consider. "Any time you touch the plant... it's the labor costs that are really hard to deal with. It's a business issue."
— Jeff Baumgartner, Site Editor, Light Reading Cable