We've written several times about "Cable's Magic Tricks"--a variety of technologies that cable operators could use to expand the capacity of the installed cable plant. Our second survey of these technologies five years ago concluded "All of these new technologies suggest that MSOs will be able to wring a lot more bandwidth -- probably many gigabits per second -- out of the existing plant. Taken together, they promise to postpone the need for additional massive capital investments for many years."
Until recently, many of these technologies seemed to be "solutions looking for a problem" since operators saw little or no reason to spend money to deploy them widely throughout their cable systems.
But the world has changed. Look at some of the recent articles and ads:
The impetus is clear--cable faces competition from multiple directions. And that competition has a lot to do with having lots of bandwidth at their disposal.
After the recent Cable Show, we wanted to know how seriously cable is responding to the competitive threats. We spoke by phone with Jeff Walker, Motorola's senior director of marketing. He confirmed that operators are on the move to expand plant capacity, and said "it's all about competition". He pointed to Verizon's FIOS service and satellite's threats of "150 HD channels" as key driving forces.
This article summarizes some of those competitive forces and looks at the wide variety of solutions available to MSOs, ranging from those that re-claim existing spectrum to those that expand it. It concludes "the sky is not falling" and that MSOs don't need to run out and put in fiber any time soon. They will have to spend some money, but nothing like what they spent to build their current cable plant.
In this article our main focus is on some of the newer and more innovative solutions. At the end, we provide links to our earlier articles which provide a more detailed exploration of the technologies that have been around for a long time, waiting for the MSOs to feel the competitive threat.
One Hundred--The Magic Number
While we don't believe in numerology, it certainly seems like 100 is today's magic number for the cable industry. That number simultaneously represents:
Here's the background on these.
Over the past fifteen years, the North American cable industry has upgraded its entire cable plant. Each major cable operator has consolidated its video and internet resources into centralized metropolitan-scale headends; built extensive redundant fiber rings throughout metropolitan areas; rebuilt the coaxial cable plant to carry frequencies up to 750 MHz or higher; and deployed optical nodes in each neighborhood to connect the fiber rings and coax plants. This upgraded hybrid fiber-coax (HFC) plant has cost the cable industry nearly $100 billion.
The North American telephone companies are becoming serious competitors to the cable operators. While the cable operators talked openly for many years about their plans to offer high-speed data and telephone service over the HFC networks, the telcos were slow to respond. Only when cable operators grabbed the lion's share of high-speed Internet service did the telcos start deploying DSL broadly and searching for ways to compete--mostly on price. When the cable companies demonstrated that they could take away a substantial number of primary residential voice customers--the telco's long-time cash cows--the telcos responded by starting to test IPTV.
Now all major telcos are rolling out the "triple play" of voice, video and data services. They are deploying more fiber to increase the capacity of their plant. Verizon is deploying fiber all the way to the home; others--including AT&T--continue to use copper twisted pair for the "last mile" to the home, much as the cable operators have continued to use copper coaxial cable.
100 Mbps Data
"100 Mbps to every home" has long been a mantra of the high-tech industry. In March of this year, the Fiber to the Home Council called for a "100 Megabit Nation" and asked the US Congress and President to adopt a strategy and timetable in 2007. On May 9th, Senator Rockefeller introduced S. Res. 191 calling for such a broadband policy. Verizon's FIOS service is based on a fiber to the home (FTTH) infrastructure capable of 100 Mbps service or more with the proper electronics at each end.
For some time now, the cable industry has been developing DOCSIS 3.0, the latest generation of cable modem technology. DOCSIS 3.0 bonds several individual cable channels--each operating at 30 to 40 Mbps--to provide peak data rates higher than 100 Mbps. Until recently, the industry had approached DOCSIS 3.0 with the same measured pace as previous cable modem standards, with years elapsing between the start of an initiative and volume production of products certified by Cable Television Labs.
To head off the perceived threat from FIOS, cable companies now want to accelerate DOCSIS 3.0 deployment. In April, CableLabs announced a "novel qualification testing approach" for cable modem termination systems "to encourage CMTS makers to submit gear for testing earlier than they otherwise might." It established Bronze, Silver and Full qualification levels to achieve "downstream data rates of 160 Mbps or higher and upstream data rates of 120 Mbps or higher". Ralph Brown, CableLabs' CTO, said he expects equipment to be submitted for testing in the fourth quarter of this year, to "enable our members to compete effectively in the very high-speed broadband data marketplace."
Comcast, the largest North American cable operator, has been one of the leaders in this effort. At the Cable Show, Comcast's CEO Brian Roberts demonstrated downloading a file at 150 Mbps using DOCSIS 3.0 over four bonded channels--he called this "Wideband" compared with today's "Broadband".
"Better than 100 Mbps" is becoming the new cable mantra. But most cable systems don't have spare bandwidth to assign for four (or more) bonded channels.
100 HDTV Channels
In the US, the direct-to-home satellite companies -- DirecTV and EchoStar -- offer the only nation-wide competition for residential subscription video, cable's cash cow. Since satellite is a comparatively poor way to provide data and voice service, the satellite companies have concentrated on high-quality video--with some popular services carried exclusively on satellite--and have taken a significant share of cable's video customers.
The latest competition is in high-definition television. After years of waiting, high definition TV sets (HDTVs) are now flying off retailer shelves and prices are dropping fast. The Consumer Electronics Association (CEA) reports that 28% of US homes had HDTVs by the end of 2006. While some owners receive the HD signal free off-air, most pay to get it from a cable or satellite set top box.
At CES in January, DirecTV announced that its latest satellites will give it the capability to carry 150 channels of HD, and said that by the end of 2007 it plans to have 100 operating HD channels--far more than are now available on any cable system. DirecTV has recently been running ads encouraging cable subscribers with new HD sets to switch to satellite to receive these promised HD channels.
While analysts have expressed doubt that there will really be 100 distinct HD satellite channels at yearend, and whether the new channels will add much value, they agree that the number of channels available in HD will increase markedly this year, and that cable operators will need to increase the number of HD channels to compete successfully for the high-end customers who subscribe to HD services.
As new HD channels find a place on satellite, they are trying to sign carriage agreements with cable operators. But HD needs a lot of scarce cable bandwidth.
Other New Services
The demand for additional cable capacity is driven by more than pressure from satellite and telephone competitors.
Overcoming Channel Lock
The cabling and electronics for a modern HFC cable plant have been engineered to provide a specific spectrum for services. The coax portion of the plant typically allocates the spectrum from 50 to 750 or 860 MHz to carry signals "downstream" from an optical node in each neighborhood to nearby customer homes, and allocates 5 to 42 MHz to carry signals "upstream" from homes back to the node. The fiber portion of the plant typically uses two fibers, one in each direction, to carry the same spectrum between the centralized cable headend and optical nodes. The downstream spectrum is divided into fixed 6 MHz channels (in North American systems), with more flexibility upstream.
The cable plant was originally engineered to carry analog television programming, and most systems still assign most downstream channels to analog programs--with the balance split among digital cable, HD, video on demand, high-speed data, and telephone service. Many if not most systems are "channel locked", with nearly all available channels assigned to services.
Cable operators clearly need to assign more capacity for HD and VOD video programming, DOCSIS 3.0 data, new telephone customers, and business services--but where will they find the space?
Broadly speaking, operators can use two approaches to gain capacity: make better use of their existing installed cable spectrum, or expanding the spectrum. Operators are likely to invest in reclaiming existing spectrum before turning to expanding spectrum, which usually requires substantially larger front-end capital expenditures.
Making Better Use of Existing Spectrum
Operators have a wide variety of choices to reallocate the existing spectrum. Since most of the spectrum is still allocated to analog programming, reducing or eliminating analog channels represents their best opportunity to recover large swathes of spectrum, but they need to figure out how best to address the huge installed base of analog-only TV sets. Switched digital video and DOCSIS 3.0 would allow more efficient use of the existing spectrum.
Going "All Digital"
The simplest solution would be to go "all digital" and eliminate analog programming as quickly as possible. Since the 6 MHz space of one analog channel can hold ten or more standard-definition digital channels, converting all analog programs to digital would free up an enormous amount of capacity that could be reassigned for other services. Most operators have been gradually reducing their analog programming in favor of digital, and they could accelerate it.
Analog programs work on old-fashioned "cable ready" TVs without a cable box. North American homes have a huge installed base of these, used without cable boxes as second, third and fourth TVs; other devices such as VCRs and PVRs also have analog-only "cable ready" tuners. (In our home, in addition to three TVs connected with set-top boxes, we have three more without cable boxes: two in guest bedrooms and another in the exercise room; we also have a VCR that's not connected through a cable box.) Since customers are unwilling to give up these legacy "cable-ready" TV sets and other devices, cable operators are understandably reluctant to remove the analog services until they have a way to support them.
This problem will gradually go away, since the United States is ending analog television broadcasting in February 2009, less than two years from now. Starting last month, the FCC has required retailers to display a label with the following text on or adjacent to any analog-only TV offered for sale: "CONSUMER ALERT This television receiver has only an analog broadcast tuner and will require a converter box after February 17, 2009, to receive over-the-air broadcasts with an antenna because of the Nation’s transition to digital broadcasting. Analog-only TVs should continue to work as before with cable and satellite TV services, gaming consoles, VCRs, DVD players, and similar products. ..." This is expected to accelerate the adoption of digital television sets and reduce the number of analog sets still in use.
Comcast has been aggressively moving toward an "all digital" service. Multichannel News reports that Comcast recently notified customers in Chicago that on July 1 it will stop carrying most analog channels. It will remove 38 channels from its expanded-basic tier, leaving only 34 channels in basic analog--the "off air" broadcast channels and a few cable channels.
But Comcast customers will need a digital set top box for each "cable-ready" TV set to receive the expanded-basic channels. They'll have to pay monthly lease charges for boxes on additional TVs, and they'll need to find a place to put the box. And Comcast will have to buy lots of cable boxes that become boat anchors when the analog TVs fade away.
BroadLogic -- How To Avoid All Those Boxes
BroadLogic Network Technologies--a Silicon Valley startup which includes Cisco, Intel and Time Warner as investors--proposes a different approach: don't buy any of those cable boxes. At the Cable Show, we met with BroadLogic's Danial Faizullabhoy, President and CEO, Jeff Huppertz, VP of Marketing & Business Development, and Del Allison, VP of Worldwide Sales.
Danial and Jeff described BroadLogic's innovative approach: broadcast all the channels in digital--but instead of using a set top box for each "cable-ready" device, install a small box just outside the home that converts the entire digital lineup to analog. That way, all "cable-ready" devices will keep working without any changes inside the house.
At first blush, this would appear to be a very costly approach, but BroadLogic's innovative chips make it possible. Their BL12000 TeraPHY wideband receiver (shown on the left) can simultaneously tune and demodulate sixteen 6 MHz cable channels. Their BL80000 TeraPIX (on the right) is a massively parallel video processor that can convert a complete digital cable lineup to analog. Packaged together in a box, these two chips can tune and demodulate the entire digital 80-channel extended-basic channel lineup, and convert it back to analog to serve all the analog tuners in the house.
BroadLogic's chips can be used in other cable applications. A chart in their booth showed many applications in the headend, in the distribution plant, and in the home--both for video and data applications. For example, the BL12000 can be used to create a cable modem which BroadLogic says "can receive over 600Mbps of downstream data."
Switched Digital Video--Making more effective use of channels
Whether analog or digital, all broadcast television channels are subject to Pareto's famous "80/20 law"--in this case, predicting that 80% of the viewers watch 20% of the channels. At any given time, many channels are not being watched by anyone in some segments of the cable distribution plant.
Switched digital video (SDV) provides the technology to take advantage of Pareto's Law by broadcasting only those programs that are actually being watched. All the less-popular programming is moved from fixed channel slots to a smaller number of variable slots; the SDV equipment turns on the channel for a specific optical node only when someone in that node tunes to it.
SDV has been around for a long time. We reminded ourselves just how long by pulling an old conference T-shirt out of Dave's drawer. The shirt says: "5th Switched Digital Video Conference" and the date is September 1995! More recently, BigBand Networks was the pioneer of the current generation of SDV--we wrote about BigBand and what was then called "switched broadcast services" five years ago. Now that cable operators are deploying SDV, other companies also offer SDV equipment.
SDV can be used for both standard-definition and high-definition video. It is likely to be particularly valuable as new HD channels come onstream, since many of these are likely to be rarely watched in the early days of HD.
Expand Available Spectrum
Just as cable operators have many ways to make better use of the existing spectrum, they also have many ways to expand the available spectrum. Over the past fifteen years, they spent nearly $100 billion to expand their spectrum (to an upper limit of 750 or 860 MHz) by physically replacing most of the original wiring and electronics in the cable plant. Further expansion (typically to 1 GHz) is possible, and some systems were engineered to facilitate this approach. Operators have also long planned to extend fiber deeper--reducing the number of homes per node to 125 or even 64--allowing reuse of the upstream and downstream channels allocated to VOD, high-speed data, and telephone service.
Some analysts have recently argued that to compete with FIOS cable operators now need to replace their HFC plant with fiber to the home (FTTH) and move toward IP video distribution--at a cost comparable to, if not higher than, the $100 billion or so they spent to build the HFC plant. While we have long believed that FTTH makes a great deal of sense for greenfield construction, we think existing HFC networks have many more years of life.
"Spectrum Overlay" from Vyyo and PhyFlex
Both Vyyo (formerly Xtend Networks) and PhyFlex Networks (formerly Narad Networks) have long promoted "spectrum overlay" techniques that carry additional services over the existing coaxial cables while operating at frequencies above 1 GHz. Vyyo provides additional RF channels, which PhyFlex provides Carrier Ethernet for IP services.
Vyyo's system creates additional 6 MHz channels which can be assigned to services the same way as the current channels below 1 GHz. Vyyo has written a white paper Preparing for an unknown future comparing its spectrum overlay approach with other techniques to expand spectrum. Not surprisingly, the paper concludes that spectrum overlay would cost operators substantially less than FTTH or extension to 1 GHz over deeper fiber.
In March, Vyyo announced that StarHub, a Singapore telecommunications company and cable operator, would use a combination of Vyyo's 3 GHz spectrum overlay and DOCSIS 3.0 to meet the challenge of Singapore's Next Generation National Broadband Network (Next Gen NBN) project. Vyyo says this will be built on StarHub's existing HFC infrastructure, and will double the downstream channel capacity and quadruple the upstream capacity.
PhyFlex's approach uses Ethernet over fiber and coax to carry IP services. PhyFlex recently changed its name to emphasize the flexibility of its approach, which can provide 1 and 10 Gbps Ethernet over fiber, and 100 Mbps Ethernet over coax. With many cable operators placing a strong emphasis on services for businesses, PhyFlex now positions its solutions to allow operators to deliver commercial-grade voice and data services over their existing fiber and HFC plant. PhyFlex recently announced that it has received Metro Ethernet Forum (MEF) certification for its Carrier Ethernet products, and claims to be the first to be able to provide certified Carrier Ethernet services for 100 Mbps Ethernet over coax.
More colors on fiber
Coaxial cable is not the only potential bottleneck in the HFC plant. As more services are delivered to homes and businesses, operators start running out of capacity in their fiber rings. At one session at the Cable Show, several papers discussed solutions for expanding the capacity of fiber rings.
We were particularly struck by the paper "Multi Wavelength Access Networks: A Practical Guide to Implementation" by Ray Thomas1 of Time Warner Cable and Venk Mutalik of C-COR. They observed that the typical HFC fiber node is served by two fibers, one in each direction, each carrying one wavelength or color. They then worked through a series of techniques to use the existing fiber to carry progressively more bandwidth, culminating with a single fiber carrying ten wavelengths, some forward and some reverse, and including two symmetrical channels of Gigabit Ethernet (with 10 Gigabit Ethernet coming soon). They concluded that the economics of multi-wavelength optics technology makes it "an attractive alternative to construction of additional fiber for increasing throughput capability".
Cable has just started exploiting many of these technologies to pull more bandwidth out of their existing plant, and should be able to keep doing so for many years to come. They'll have to spend some money to do so, but it won't be anything like another $100 billion.
MSOs can proudly point to their many positive results, as they did at the Cable Show. However, in a competitive world, nothing stands still. Irving Berlin said "The toughest thing about being a success is that you've got to keep on being a success."
For further reference:
( www.motorola.com ) ( www.ftthcouncil.org ) ( www.cablelabs.com ) ( www.comcast.com ) ( www.ce.org ) ( www.directv.com ) ( www.broadlogic.com ) ( www.bigbandnet.com ) ( www.vyyo.com ) ( www.phyflex.com ) ( www.starhub.com ) ( www.timewarnercable.com ) ( www.c-cor.com )