Despite pouring more than $100 billion into upgrading their hybrid fiber-coax (HFC) networks over the past decade, North American cable operators could soon find themselves outgunned by Verizon Communications Inc. (NYSE: VZ)'s emerging fiber-to-the-premises (FTTP) network, according to a new report.
The latest Cable Industry Insider report from Light Reading compares the competing technological visions of the cable industry and its most threatening telco rival. Most notably, it concludes that Verizon's new FiOS service will likely provide much more shared bandwidth than even the cable industry's new Docsis 3.0 technical standard from Cable Television Laboratories Inc. (CableLabs) .
"By most accounts, FiOS wins the bandwidth battle against Docsis 3.0 hands down, a point that even the most partisan cable engineers and strategists usually concede," the report says. "Using its existing BPON architecture, Verizon is able to offer shared bandwidth of 622 Mbit/s downstream and 155 Mbit/s upstream, divided among no more than 32 homes."
By the same token, the report concludes that Verizon will probably deliver much faster broadband speeds with FiOS than cable operators could ever hope to match, even with the adoption of the Docsis 3.0 specifications. Although the new Docsis standard will enable MSOs to offer shared data speeds as high as 160 Mbit/s downstream and 120 Mbit/s upstream, it says, FiOS has the capability to offer even higher speeds because of its superior network capacity.
"Thanks to a clear advantage on the overall bandwidth front, FiOS easily beats cable's best in the data speed department," the report states. "With the introduction of Docsis 3.0, cable operators will be able to deliver far greater speeds to customers than they do today. But, due to the basic bandwidth gap, they still won't be able to offer the same potential speeds as FiOS."
Highlighting another key metric in FiOS's favor, the report finds that Verizon's growing FTTP network will cost less to operate and maintain than the cable's industry's comparative hybrid fiber-coax (HFC) plant because of the telco's reliance on passive optical networking (PON) technology. With PON, an outside plant requires no electronics, unlike the cable industry's plant.
In one other advantage for the Verizon service, FiOS has a big headstart on Docsis 3.0 on the deployment front, according to the report. Plans call for the FiOS suite of services to pass 6 million homes by the end of this year and 12 million by the close of 2008. In contrast, the report says, cable operators probably won't be able to start introducing Docsis 3.0 systems and gear en masse
until 2008 at the earliest.
But don't count out Docsis 3.0 just yet, the report says. Once Docsis 3.0 products finally start churning off the assembly lines, it states, cable operators should quickly begin to catch up to Verizon because no new cable plant need be built to deploy the new equipment.
"Docsis 3.0 will undoubtedly be almost ubiquitous in cable systems by 2010 or 2011, just as Docsis 2.0 is today," the report says. "Thanks to this rollout and cable's huge installed base, which is at least twice as large as Verizon's, Docsis 3.0 will end up reaching far more homes than FiOS over the next few years."
Also in the cable industry's favor, the report asserts that Docsis 3.0 should run up much lower deployment, connection, equipment, and services expenses for MSOs than FiOS is already racking up for Verizon. For example, Verizon reports that it now costs more than $1,800 to serve each FiOS customer -- far more than it should cost cable operators to serve each Docsis 3.0 subscriber, because of the great expense of building the new FTTP network and installing set-top boxes, modems, and other equipment in every FiOS home.
In addition, the report gives a big nod to cable in the installed plant category. With a current base of 65.5 million basic subscribers, it notes, cable operators reach far too many homes for Verizon to ever catch up, short of merging with rival AT&T Inc. (NYSE: T).
The report ranks the two industries evenly on the services that their new technologies enable, at least for the moment. But it points out this category could shift in Verizon's favor as more bandwidth-intensive applications evolve.
"Docsis 3.0 vs. FiOS: Tale of the Tape" is available as part of an annual subscription (six monthly issues) to Light Reading's Cable Industry Insider, priced at $1,295. Individual reports are available for $900.
So what you're really saying is that you're running Equal Cost MultiPath. That doesn't give you a true 80M upstream pipe since any given flow has to always live on the same upstream. DOCSIS 3.0 channel bonding maintains packet ordering. Your method doesn't. You can't run, say, a file transfer test and get 80M. You'll instead typically bump into the DOCSIS 1.1 channel width limit (3200 KHz in 1.1 6400 KHz in 2.0) on an upstream and the reality that the upstream plant isn't clean enough to run better than QPSK. On a sunny day, you'll get 10 meg upstreams out of Docsis 1.1 with 16QAM and 20 meg on DOCSIS 2.0 by running 64000 KHz channel width if you've gotten all the 1.1 CPE off that frequency. It's usually much worse than that. Have you ever seen real plant in the field?
A quick review of some of the documents at www.cablelabs.com might help you understand better how these systems operate.
So I don't need my DOCSIS Doczone account? That's a relief.
Upstream? I'm only aware of deployments with downstream channel bonding. I believe you need DOCSIS 3.0 silicon to bond upstreams.
Who said they were bonded. There are multiple channels with modems distributed amongst those channels. Modems can be statically or dynamically configured to "tune" to different frequencies. Bonding is only necessary when provisioned service approaches or exceeds channel capacity. Depended on configuration, upstream channel capacity is between 5-30M but more routinely in the 10M range.
An 80M upstream would destroy the MSO IP network with a storm of BitTorrent P2P traffic. That might be a good thing for Cisco since the MSOs would have to forklift all their routers. For a business that needs that kind of bandwidth, it makes more sense to drag a dark fiber to them.
Each cable modem has a provisioned speed and therefore cannot use the total aggregate capacity. Cable Modems have QoS queuing capabilities that are set at activation time and can be modified dynamically. Im sure you have noticed the advertizsd upstream and downstream speeds.
Sorry, no storms. A quick review of some of the documents at www.cablelabs.com might help you understand better how these systems operate.
ozip writes: I know of a US operator which has 80M of upstream bandwidth capacity using multiple channels today.
Upstream? I'm only aware of deployments with downstream channel bonding. I believe you need DOCSIS 3.0 silicon to bond upstreams.
An 80M upstream would destroy the MSO IP network with a storm of BitTorrent P2P traffic. That might be a good thing for Cisco since the MSOs would have to forklift all their routers. For a business that needs that kind of bandwidth, it makes more sense to drag a dark fiber to them.
Selling using speed results in attracting the unprofitable subscribers. I havent seen numbers from FIOS but Im willing to bet that they have more than their "fair share" of heavy use subscribers.
I just think this whole discussion is nothing more than hype. Cable operators are working on mechanisms that will free downstream channels and in the future the combination of DOCSIS 3.0 and IPTV will help. One of the most overlooked but important additions to DOCSIS 3.0 is a dramatic improvement in multicast functionality necessary for IPTV development. Also, as I have stated before, I know of a US operator which has 80M of upstream bandwidth capacity using multiple channels today.
The CMTS and cable modem vendors and their silicon suppliers aren't in a huge hurry to invest tens of millions in R&D for features in DOCSIS 3.0 that won't be deployed. Vendors are finally making some money off all of that DOCSIS 2.0 and PacketCable 1.0 investment. I don't see many companies willing to pump hundreds of millions of dollars into DOCSIS 3.0 and PacketCable 2.0 development.
True in principle. Except, this is exactly what silicon, CPE and CMTS vendors have done with DOCSIS and PacketCable for every new spec release over the last 10 years. MSOs pressure them to deliver solutions complying with the latest spec. The vendors deliver, the MSOs deploy it, but do not turn on the functionality of the newest spec for years. MSOs just don't want to bear the risk of hardware obsolescence and place it on the backs of vendors. For some reason, the vendors agree to carry the burden. The same thing will happen with DOCSIS 3.0. It's a tough cycle to break.
schmitt writes: The problem is that the Cable plant lacks extensibility on the UPLOAD side. If an application emerged that really placed a premium on fast upload speeds, Cable would be toast, and would have to sink some serious $$$ to compete.
The MSOs are already seeing substantial loads on their networks 24x7 from P2P media stealing applications like BitTorrent and Slingbox. The last thing they want to do is further clog their networks with this traffic by having bigger upstream pipes. This is the whole DOCSIS 3.0 vs DOCSIS 2.0b argument. Who would ever want channel bonding on the upstream? Today's DOCSIS 2.0 silicon can already support some amount of downstream channel bonding and IPv6. The CMTS and cable modem vendors and their silicon suppliers aren't in a huge hurry to invest tens of millions in R&D for features in DOCSIS 3.0 that won't be deployed. Vendors are finally making some money off all of that DOCSIS 2.0 and PacketCable 1.0 investment. I don't see many companies willing to pump hundreds of millions of dollars into DOCSIS 3.0 and PacketCable 2.0 development.
Given the current set of applications that home users demand, there is no discernible difference between FTTH and Cable, unless you do P2P which I characterize as a fringe technology.
Agreed. In the near term, this is primarily about bandwidth bragging rights for marketing purposes, i.e. one side being able to say their service is "fastest." Ironically, it's cable's own silly speed-centric product positioning strategy coming home to roost. Both cable and FTTH are poised to provide the Internet bandwidth consumers genuinely need and use for quite some time. Unless, as you say, it becomes an upstream-centric world.
Given the current set of applications that home users demand, there is no discernible difference between FTTH and Cable, unless you do P2P which I characterize as a fringe technology.
What will help the cablecos more than DOCSIS 3.0 (which is really just channel bonding) is the conversion of Analog video spectrum to digital. Lightreading covered this with the Broadlogic product announcement. I talked about it here. This will really open up spectrum for more bandwidth and allow the DOCSIS 3.0 channel bonding to really be utilized.
The problem is that the Cable plant lacks extensibility on the UPLOAD side. If an application emerged that really placed a premium on fast upload speeds, Cable would be toast, and would have to sink some serious $$$ to compete.
Comparisons of nailed-up bandwidth capacity are misleading. So FIOS would be great at delivering unicast high-bandwidth constant-bit-rate services that have a 100% peak usage rate. So what? In real life, services that get delivered to residential customers are bursty; or are multicast; or have a peak usage rates of well under 100%. In other words, amenable to stat-muxing. Cable architecture excels at delivering that type of service.
The blogs and comments are the opinions only of the writers and do not reflect the views of Light Reading. They are no substitute for your own research and should not be relied upon for trading or any other purpose.
To save this item to your list of favorite Light Reading content so you can find it later in your Profile page, click the "Save It" button next to the item.
If you found this interesting or useful, please use the links to the services below to share it with other readers. You will need a free account with each service to share an item via that service.
Digg
Del.icio.us
Reddit
