Waiting for Docsis 3.0

Will 2008 be the cable industry's big kickoff year for Docsis 3.0?

Judging from the recent pronouncements by several major North American MSOs, there's good reason to believe it might. In its earnings call last month, Comcast Corp. (Nasdaq: CMCSA, CMCSK) affirmed plans to roll out the next-gen broadband spec to 20 percent of its footprint by year's end. Both Charter Communications Inc. and Mediacom Communications Corp. said they plan to start testing Docsis 3.0 equipment later this year. And Vidéotron Telecom Ltd. , one of Canada's biggest cable operators, unveiled a new ultra-fast, wideband service that uses "pre-Docsis 3.0" technology.

Yet, as we'll explore in "Docsis 3.0 Strategies," a Light Reading live event slated for March 19 in Denver, the answer to the question appears to be more of a decided maybe. While some major cable operators around the world are scrambling to deploy Docsis 3.0 as soon as possible, others are taking their own sweet time.

Take Time Warner Cable Inc. (NYSE: TWC) and Cox Communications Inc. Unlike Comcast, neither large U.S. MSO is aggressively pushing to launch Docsis 3.0 in 2008. Both seem content to wait until sometime next year to introduce the channel-bonding spec, which will enable cable operators to offer shared speeds of more than 100 Mbit/s upstream and downstream. As Jay Rolls, senior vice president of technology for Cox, put it at an industry forum last year: "We're not dying for this; we don't need it tomorrow."

Or take Rogers Communications Inc. (NYSE: RG; Toronto: RCI). Unlike the bolder Videotron, Canada's largest MSO has adopted a wait-and-see attitude about Docsis 3.0. Like Time Warner and Cox, Rogers doesn't plan to begin rolling out the next-gen spec until 2009, at the earliest.

The same contrasts in MSO strategies can be seen elsewhere in the world, even within the same companies. In Japan, for example, Liberty Global Inc. (Nasdaq: LBTY)'s Jupiter Telecommunications Co. Ltd. (J:COM) has already signed up 17,000 subscribers for its new wideband, pre-Docsis 3.0 service, which offers shared download speeds of up to 160 Mbit/s. But, in its western European and Latin American markets, the international MSO has no intention of introducing such blazing speeds any time soon.

Why such a disparity? The chief reason why some cable operators are embracing Docsis 3.0 so tightly and others are not is the state of telco competition. In the U.S., such early Docsis 3.0 champions as Comcast are facing strong challenges from Verizon Communications Inc. (NYSE: VZ)'s growing fiber-to-the-home (FTTH) network. Verizon, which tangles with Comcast up and down the East Coast, has now signed up more than 1 million FiOS TV customers and more than 1.5 million FiOS Internet users.

Similarly, in Japan, J:Com is fighting an uphill battle against the entrenched FTTH networks of the incumbent phone companies. So the MSO, like Comcast and others, is turning to Docsis 3.0 as a way to match the soaring transmission speeds of its rivals.

But the threat of stiff competition is not prompting all MSOs to move so fast. In the New York area, for example, Cablevision Systems Corp. (NYSE: CVC) doesn't aim to deploy Docsis 3.0 until at least the end of the year, even though it arguably faces greater FiOS competition from Verizon than anybody else.

Clearly, there are other factors at play too. Some MSOs are not racing to test and deploy Docsis 3.0 because, well, there simply isn't all that much equipment to test and deploy right now. While Cable Television Laboratories Inc. (CableLabs) has approved cable modem termination systems (CMTSs) from three vendors -- Arris Group Inc. (Nasdaq: ARRS), Casa Systems Inc. , and Cisco Systems Inc. (Nasdaq: CSCO) -- so far, it has not yet certified any cable data or voice modems. So cable operators must either take their chances with gear that hasn't been approved or wait until the first certified modems pass muster this spring or summer.

At the same time, other big MSOs, such as Cablevision, are proceeding cautiously because they don't see the crying consumer need for Docsis 3.0's high speeds just yet. Confident that they can deliver download speeds of 30 Mbit/s or more with their existing Docsis 2.0 networks, they believe that can match any telco offering for the time being.

Indeed, even Videotron, the North American poster child for Docsis 3.0, isn't using its new channel-bonding, wideband architecture to offer shared downstream speeds higher than 30 Mbit/s and 50 Mbit/s. Nor does it plan to do so in the near future, even though it successfully tested download speeds of 100 Mbit/s and greater in beta trials with Cisco.

"We want to gear our services to the majority," Manon Brouillette, Videotron's senior vice president of marketing, content, and product development, told analysts last month. She noted that Videotron officials "determined that speeds of 30 Mbit/s and 50 Mbit/s suited the needs of the market, as well as consumer demand, existing technological capabilities, and the Internet chain."

So cable's Docsis 3.0 era may be starting this year. But it's going to be a while before it firmly takes hold.

— Alan Breznick, Senior Analyst, Heavy Reading

alchemy 12/5/2012 | 3:46:03 PM
re: Waiting for Docsis 3.0 The reality is that the RF plant is out of spectrum. Downstream and upstream channel bonding may sound great but end users will not see any improvements in performance until the MSOs either do expensive node splits to have less subscribers sharing the spectrum or until the MSOs reclaim bandwidth used elsewhere (analog television channels is the obvious place).

High Def video is the cash cow where the MSOs attract the highest yield customers. It stands to reason that they will temporarily allocate their scarse bandwidth to HTDV rather than improving residential data service. Commercial data is going to use fiber to the customer with Metro Ethernet and won't need DOCSIS 3.0 other than areas where it is uneconomical to drag a fiber to the customer.

I suspect Comcast is hot on IPv6 since they're short on IP address space. Their DOCSIS 3.0 love is for one feature which could be made available in today's DOCSIS 2.0 equipment with a software upgrade. DOCSIS 3.0 CPE gear is expensive since you need to drive multiple channels to support channel bonding. It's no longer a $10 parts cost cable modem. If it's more expensive, nobody is going to buy it... thus the slow availablity of DOCSIS 3.0 CPE.

I can see upgrading the CMTSs to DOCSIS 3.0 and migrating to higher density upstream-only and downstream-only CMTS blades. You get more density in your equipment rack, you get more flexibility in your upstream vs downstream mix, and your price per bit at the CMTS drops since the blades are denser. I don't see wide deployment of 3.0 in the CPE gear anytime soon since it's a big cost with no benefit for the residential market. If they're already traffic shaping the upstream to limit P2P traffic, I don't see upstream channel bonding ever being deployed. If your plant is currently at DOCSIS 1.1, it makes complete sense to upgrade your CMTSs to DOCSIS 3.0 and leapfrog 2.0.

Bonding also induces latency since the packets can arrive out of order and have to be reassembled. When you have dropped packets, you end up with a ton of jitter since you're waiting for a packet that didn't arrive. You can't use it for true real time traffic like voice so it's useless in PacketCable 1 VoIP deployments. Commercial traffic is also sensitive to latency and jitter. If you're competing against FIOS, the last thing you want is for Verizon to publish the lousy DOCSIS 3.0 jitter and latency numbers compared to fiber.
Luke M 12/5/2012 | 3:46:02 PM
re: Waiting for Docsis 3.0 "High Def video is the cash cow where the MSOs attract the highest yield customers. It stands to reason that they will temporarily allocate their scarse bandwidth to HTDV rather than improving residential data service."

Data (and voice) are by far the most profitable services on a per-Mhz basis.
robrob1850 12/5/2012 | 3:45:51 PM
re: Waiting for Docsis 3.0 I believe that our biggest challenges are going to come in the form of upstream speed issues. Unfortunately most cable operator's HFC plants were largely built for downsteam. Achieving DOCSIS 3.0 upstream speeds will require higher upstream modulation profiles for the CMs and the CMTS, 64 QAM and ATDMA becoming the norm. I is going to be time for cable operators to re-evaluate their HFC plants, especially given the fact that a 90 degree bend in a coax cable can be disruptive in a 64 QAM environment. I also think we should be looking at our combiners and passives; the upstream spectrum has always been a particularly nasty place to do business and if we want to make it more effective and bond these channels at 30 Mbps each, we need to start from the customer premise and then work forward.
Sign In