Cable's eventual migration to an all-digital environment is not a smooth, well-paved, one-way road, but a tangle of different paths that could wind circuitously, yet lead to the same destination. Some corridors are easier than others, while other passageways are more trying and may take longer, but come with a bigger payoff.

The options on how to get there are many, but what is much clearer are the benefits such a migration could provide to the cable industry.

The most obvious benefit is the reclamation of analog spectrum at a fraction of the cost of the original two-way upgrades to 750 MHz or more. Depending on the system and operator, estimates are that about 75 percent of that bandwidth is used for analog video.

Digital bandwidth is also much more valuable than its analog forefather. Operators, still under siege by competitors such as DBS, can leverage it to flesh out their HDTV lineup or use it to bump up their VOD offerings.

But seeing where one needs to go, and knowing how to get there, are completely different things. Still, MSOs and vendors alike are already trying to figure out the how.

One option–rebuilding again–is not really an option. Cable operators "aren't going to paint the bridge again and go to 860 MHz," says Mark Davis, vice president of professional services for BigBand Networks. "They can't afford it, and the Street is not going to let them spend that kind of money again."

Digital tools and schools

The 2003 National Show in Chicago showcased the first official all-digital "hows" to emerge. Pace Micro Technology introduced the Digital Cable Adapter (DCA), a two-tuner "set-back" device that will likely go for $69.

Motorola Broadband also unveiled the DCT700, a "sub-$100" all-digital box.

Despite some apparent strengths, each approach wasn't without its critics. Motorola's box preserves the company's proprietary conditional access (CA) platform.

The version Pace showed off does not have HDTV in its repertoire. Some operators also wondered more privately about how and if it will comply with EAS. But an all-out plunge by cable operators into the all-digital sea is not a here-and-now concern, so there's still time to address any of those apparent shortcomings.

New and forthcoming products aside, under discussion are three primary ways for cable to go all-digital. An operator can transmit video over the HFC plant via MPEG-2 transport streams the way it's done today, or use the same plant to transmit MPEG-encoded streams, but transport them over DOCSIS. A third approach is a hybrid of the two.

"Both approaches have been proposed by MSOs and the vendors," says Tom Cloonan, chief technology officer of Arris Broadband.

Plus, no single approach will work uniformly for all MSOs. "Nobody's wrapped [going all-digital] in a package. Every operator has a different architecture, philosophy and products," adds Ben Stanger, senior manager of broadband marketing for Cisco Systems' broadband edge/mid-range routing business unit.

One approach, which wasn't a formal star on the National Show floor, involves the DOCSIS cable modem and IP-based distribution of broadcast video. Video over DOCSIS appears to be a leading candidate for the domestic cable industry's foray into all-digital.

One reason: Distributing video over DOCSIS also makes the cable infrastructure less complicated. Instead of one silo for data and IP voice and another for video, the so-called "triple play" would come under an umbrella, packet-based silo. "The key desire is to replicate the tremendous success with DOCSIS and put it toward the video network," Stanger says.

Using the DOCSIS modem as the main conduit for voice, video and data, notes an engineer with a top U.S. MSO, holds many other advantages: interactivity, an attractive price point and the semblance of a conditional access platform open to a variety of vendors. "The good thing about a cable modem is that it's under $50, and it's two-way," he explains.

Plus, thanks to BPI+ (Baseline Privacy Plus), the cable modem also contains the makings of a video CA system. BPI+, though still in need of some video CA features, offers a digital certificate mechanism that is, according to some MSO engineers, more elaborate than legacy digital cable CA systems. It's also virtually impossible to clone a BPI+-enabled device.

Now we've got both camps with a starting point of conditional access; "That's another thing we'd like to bring together," says an MSO engineering exec. "What we need out of the set-top CA is the price of BPI+: ie., practically nothing." But BPI+ still needs an entire video authorization system–one that is secure enough to placate Hollywood.

For all the good a cable modem offers, it does have some shortcomings. For one, it just passes bits through and doesn't decompress and display them on a TV set. But that could change with the addition of MPEG-2 decompression and processing technology.

That's not to say that traditional set-tops don't do plenty of good things. In addition to handling TV signals, they display on-screen graphics and interactive guides and house a CA system to authorize premium channels, pay-per-view and video-on-demand buys.

But legacy set-tops still have an "anemic" return path, as one cable engineer puts it, and they haven't come close to breaching the $50 unit barrier.

That doesn't mean the low-end boxes can't be made to be cheaper by shaving some technology off the top. Instead of 3D graphics and millions of colors, maybe 2D and thousands of colors could suffice.

Other elements that could be tossed overboard in pursuit of lower price points: the clock, S-Video connectors and superfluous buttons. Low-end box IPGs may not need to support scaled video. These are just a few of the elements under consideration by some of the largest U.S. MSOs.

One of the problems with legacy boxes–a limited return path–is already being addressed by CableLabs via the DOCSIS Set-top Gateway (DSG). The spec, already approved, boosts the box's signaling return path for interactive programming guides and other data. It's expected to give set-tops a raw upstream throughput improvement of roughly 10-to-1. Plus, it can be used alongside the existing CA system.

"DSG is one of the first steps toward combining DOCSIS and set-top boxes," Cloonan says. "It's definitely going to be out there in about two years."

If video-over-DOCSIS is the answer, the cable modem termination system (CMTS) will have to contain more downstream channel capability.

Arris' flagship CMTS, the Cadant C4, already has 32 downstream channels. "In the future, you could see 50 to you get into the video space," Cloonan says.

Although a spec like DOCSIS 2.0 preaches upstream/ downstream symmetry, the video-capable CMTS of the future will require enough flexibility to accommodate downstream-dominated streams, Cloonan explains.

But using DOCSIS to distribute video is not a path set in stone. The CMTS for video can also add up. "Maybe I've saved $100 in the set-top, but I just spent $100 per set-top on the headend side. That's not a great [cost] answer...You're just shoving the pea around," says an engineer with a top U.S. MSO.

The sub-$50 all-digital set-top isn't going to hit a home run until it deals with a few curveballs the cable industry throws its way. MSOs that are inquiring about it today are discussing palatable price points, but are also requiring that it handle an advanced video compression system. Future-proofing for an advanced codec such as MPEG-4 may call for set-tops to house a modular slot for a new chipset or a software-upgradable chipset.

"If I'm going 100 percent digital all across the country, I want to get this right," says an MSO engineer who's looking over the current options very carefully. "I don't want to roll out the wrong platform. We would've loved to launch MPEG-4 10 years ago when MPEG-2 was state-of-the-art. The next window is the 100 percent digital box. I can't miss that window."

Because the licensing terms for MPEG-4 are still questionable, some operators are also considering the use of the proprietary codecs from the likes of RealNetworks or Microsoft Corp.

Avoiding the dumb pipe

But adding video to the IP mix could introduce some risk to the cable industry's way of doing business.

Today, cable operators have control of the video services offered to the set-top box. Offering video through IP-based set-tops could transform cable into a common carrier, or "dumb pipe," enabling others to offer services over the infrastructure without the operator's control or consent.

Early indications of that have already emerged in the VoIP arena. Companies such as Vonage don't necessarily need the cable operator's blessing to offer services through the existing cable modem connection, though they are exposed to potential QoS issues.

"You could have a similar occurrence in the video world in the future, where you have a service provider that may use RealNetworks-encoded video streams and Microsoft-encoded streams...that fit into cable's available pipe," Cloonan says. "MSOs need to start [partnering] with them, or it could happen."

To avoid becoming the dreaded dummy, cable operators have and continue to invest heavily in programming and content. Comcast's apparent interest in the Vivendi media empire is seen by some as a defensive move for just this reason.

The CableLabs PacketCable Multimedia specification, with its array of policy servers and underlying DOCSIS-based QoS capabilities, might also have a key role to play in the migration to all-digital, and give operators the control they want to retain.

Smooth migration

Because operators probably won't flash-cut everything to all-digital all at once, they'll have to forge a migration plan.

"It will not be a dramatic change with operators throwing everything out there to change it," Stanger says. "It will be an evolutionary change to the architecture to protect capex and to sustain customer service as they migrate."

One proposal on the table today is a hybrid set-top that allows video to travel over the DOCSIS receiver and an MPEG receiver. That combo "can support the different types of services in a transitional period where you might have both MPEG-2 transport stream traffic and DOCSIS traffic flying down the cable simultaneously," Cloonan says.

One available tool is switched broadcast, a technology that can help operators protect their base of analog customers during the transition to all-digital. "If you start to erode your analog basic base, that's a dangerous thing," notes Davis of BigBand Networks. "Switched broadcast could give you the incremental bandwidth over time as you need it, rather than doing it with a flash-cut."

As more HD and VOD are added, for example, the least-watched channels could be moved to a switched tier. Plus, switched broadcast doesn't require operators to deploy all-digital boxes by the tractor trailer load, an expensive proposition even at $100 per unit. "If there are 3.5 boxes per home, that's still a $350 investment," he says. Those boxes are also subject to truck rolls and other potential hidden costs.

"It's the perfect, scalable way...without deploying massive amounts of set-top boxes and without disrupting your whole way of doing business and doing a mass cut to digital," Davis says.

And, because it's transparent, the customer doesn't know the difference. "It still looks and feels like the rest of the programming that [MSOs] offer today," he adds. The first customer who requests a switched program may see an incremental 200-millisecond delay, but anyone else who cuts into that active stream thereafter will see no delay.

"That's because the box has already been told where the programming exists in the spectrum and how to tune to it," Davis says.

Another option cable has in its back pocket is Sony Passage, a technology that allows CA systems to live side-by-side.

But a world full of low-end digital boxes presents some challenges if customers want more interactivity. A $35 box likely will be relatively limited when it comes to the applications it can offer.

Boatloads of thin-client, all-digital boxes are what companies like ICTV Inc. have dreamed of for more than a decade. The company's Headendware platform, which is still looking for its first deployment with a major U.S. MSO, places most of the processing and power needs on the network rather than in the set-top.

"I don't think you can get there [to all-digital] without looking at centralized processing and graphics processing capabilities," says ICTV CEO Wes Hoffman.

Journey just beginning

Cable's migration to an all-digital platform, for all intents and purposes, is still years away, but some operators are starting to make some formal moves in that direction.

It's expected that Comcast could test the 100 percent digital concept in a couple of systems late this year or early next year–well before the ultimate, $35 box emerges.

Comcast has also issued a request for information (RFI) with shades of Time Warner Cable's Interactive Services Architecture (ISA). Though Comcast's "Next-Generation On-Demand Architecture" has VOD, the network PVR and everything-on-demand in mind, it also begins to approach IP video convergence by asking for details tied to streaming media platforms for the PC. Comcast is expected to follow with an RFP later this year or in early 2004.

"It's far more advanced than ISA," says an industry insider who has seen the Comcast RFI.

Although Comcast is preparing to test the concept soon, it's expected that all-digital will initially take root internationally.

One reason: U.S. operators already have an embedded base of MPEG boxes that they won't switch out right away for IP-based devices because the switch would be an extremely capital-intensive proposition. Another: Many MSOs outside of North America operate in greenfield situations, so IP-based video is one of several sensible approaches available to them.

Among some of the early activity, a consortium of Japan-based media companies (Jupiter Programming Co. Ltd., Tohokushinsha and SECOM) formed this summer with the aim of distributing multichannel video services via DSL and fiber-to-the-home technologies. The resulting company, On-Line TV, will offer linear TV programming in partnership with major ISPs in the region.

But how to get there will remain a topic of debate and trial for a few years.

"Every MSO has camps of people who think [one method] is the right way to go in the long run. There are also camps that believe it will never happen," Cloonan says.