Readers of this blog apparently believe that speed is critical when it comes to the Internet. Yesterday's post about a post-DOCSIS 3.0 strategy attracted a fair amount of comment. The basic question is, "How will cable cope with the huge bandwidth demands of streaming video and other highly consumptive apps?"
The most critical part of getting ready for "ultraband" speeds has nothing to do with our Internet business. It has to do with how we manage our legacy video service. I thought we'd discuss a road map of that challenge by answering a question raised by one of our readers.
Mark asked,
Great question. Most modern cable systems are a hybrid between conventional coaxial cable and fiber optics. And Mark, you're right, the coax definitely is more constraining, in terms of capacity, than fiber. However, the coax is typically the very last piece of the highway and your bits and streams travel very short distances over it. Coax is extremely versatile though and, because it's practically universally deployed, it's very cost effective.
The coaxial portion of modern cable networks is typically designed at a minimum capacity of 750 MHz. Cable has to continue to provide the legacy services of the past 50 years and our plant must continue to deliver analog TV signals, at least until almost all consumers' equipment have digital tuners in them. Don't confuse this cable digital upgrade with the broadcast digital transition the country is going through until June 12th -- this is a completely different subject, exclusive to cable networks.
Here's the challenge. Analog channels take up a lot more bandwidth than digital ones. Even high definition channels, which take up to as much as six times the bandwidth that standard definition digital channels take up, are two to three times more bandwidth efficient than older analog channels. That's why cable operators eventually will to convert to mostly- or all-digital formats over the next few years. In the meantime, even more bandwidth will be set aside to simulcast the analog channels in digital, allowing us to deploy digital-only set top boxes during the transition, even for our classic cable customers.
In a year or two, most cable operators will migrate their remaining analog customers to digital and will recapture the enormous bandwidth those analog channels were occupying. Before you start flaming me, this really is NOT about forcing people up to more expensive tiers of service. It's all about managing our networks efficiently so we can deliver many more new digital services to those who want them while still providing low-cost classic cable (albeit in digital format) for others, well into the future.
To put the bandwidth challenge in perspective, today most cable systems allocate around 550 MHz (out of 750 MHz) to analog TV. Everything else we do is jammed into the last 200 MHz. That includes Broadband, Phone, standard definition digital TV, high def TV, video-on-demand, and those new interactive applications people are developing for Tru2way. We are working very hard to juggle things to make room for new services and, until we recapture the majority of that analog bandwidth, we're very tight on capacity. That's why we have and will continue to migrate individual channels from analog to digital -- so we can make room for new services that many desire now, even during this temporary period of tight bandwidth.
Soon, Insight customers will hear a lot more about simulcast, and their new ability to watch our analog channels in a duplicated digital format. That will result in better pictures and sound on new, large TV sets. And it will also allow us to install digital equipment into more and more homes so, when the day arrives that we end analog transmissions, very few customers will need to do anything.
So, Mark, as you can see, cable is in a unique position to deliver supersonic Internet speeds in the next couple of years. We'll do so by deploying DOCSIS 3.0 technology over its new-found, unused (formerly analog) bandwidth. You referred to phone companies deploying fiber to the home and, although one company -- Verizon -- is doing so, they're doing it selectively in areas where they believe people can afford to pay for expensive services. For 90% of America to have access to super-fast Internet, cable's already-installed plant will do the job quickly and cost-effectively so everyone can afford it and have access to it.




Heh, 550MHz out of 750MHz is analog. Had no idea it was that much, but I guess most of those 60 or so basic channels are all analog. Converting those to digital would really give you a lot of space. Let's do the math:
2-3 HD per channel * 6 times larger than SD digital = 12-18 compression ratio (let say 15)
550MHz / 6Mhz channels ~= 92 channels
92 / 15:1 compression ratio ~= 6 channels (at 6MHz)
6 * 6Mhz = 36MHz total
So, even if we made it 50MHz (in case the ratio isn't always 15:1), that would give you an extra 500MHz to play with, which is about 83 analog channels. Those 83 channels could be:
* 1000-1500 digital channels
* 165-250 HD channels
* 10 8-channel banded DOCSIS3.0 downstreams, totaling 3 Gbps (!!!)
Or something more realistic like:
One 8-channel banded downstream (for 100Mbps service) + 150 extra HD channels + 225 extra digital channels
And this would be on top of the existing HD/digital channels. That would be pretty cool. Looking forward to it!
Posted by: sineswiper | Friday, February 27, 2009 at 01:58 PM
hey how you doing! Nice posting. I enjoyed reading it. I too run a blog on internet and traffic building, and checking out what others may have written.
Posted by: Traffic Building | Friday, March 27, 2009 at 02:02 AM
This is gonna huge, i just cant belive it that i am standing at a great blog of my life, i am really glad to have my comment here in very decent topic. thanks to webmaster.
Posted by: Acomplia | Thursday, July 23, 2009 at 10:03 AM