Dear Saul Hansell: DOCSIS 3.0 Isn't Enough, We Need Fiber

I continue to be amazed at how many respectable, knowledgeable people still think that DOCSIS 3.0 is the answer to all our broadband problems.

Take this blog post by Saul Hansell on NYTimes.com entitled: "Does Broadband Need a Stimulus?"

Let's parse through the details of the arguments he's trying to make:

"The notion of a broadband gap is hooey."
While I have my qualms with the OECD rankings I can't see how anyone can say with a straight face that other nations aren't ahead of us on the broadband curve. And then there's the glaring gap between suburban and rural communities within the US. To say there is no broadband gap is wrongheaded at best and misleading at worst.

"With new cable modem technology becoming available, 19 out of 20 American homes eventually will be able to have Internet service that is faster than any available now anywhere in the world."
First off, we can't discount the importance of reaching that 20th home; they don't deserve to be left behind and rural communities need more robust connectivity than more densely populated areas.

Secondly, a key word in this is "eventually." What does that mean? How long will it take for DOCSIS 3.0 to get everywhere? At the pace it's going now possibly decades.

Thirdly, what good is "eventually" delivering service that is faster than is available today? Can we guarantee that it will be fast enough to keep up with the demands of tomorrow?

"Running a new fiber-optic cable to every American home may well increase competition in broadband providers, but it isn't needed to deliver high-speed Internet service."
The only way we'll have competition in wireline broadband is if there's a fiber pipe and a DOCSIS 3.0-enabled copper pipe. Without fiber, cable will become a monopoly as demand for bandwidth increases and wireless can't keep up. If we want competition, we have to have fiber.

Additionally, while you may be able to deliver "high speed" Internet service, cable networks will always be less reliable than fiber. They're more prone to equipment malfunctions, and the more people using the cable network the worse its quality becomes. If we want people to rely on networked experiences more, they need reliable networks, and that means fiber.

"What is most significant about Docsis 3 is that it turns out to be quite inexpensive to upgrade existing cable systems to use it. As a result, Comcast and other cable systems are already deploying the technology rather quickly. In other words, with no government intervention, the country is going to have the infrastructure very soon to provide almost everyone with the fastest possible Internet service."
True upgrading cable networks to DOCSIS 3.0 is cheaper than laying new fiber, which is why I'm all for seeing all cable networks get this upgrade. But to say this technology is being deployed "rather quickly" is rather disingenuous. Only some cable operators are deploying it at all, and most of those are only doing so in areas where they face competition from fiber. It'll be many years before DOCSIS is anywhere near universally available.

And to say that DOCSIS 3.0 will deliver "the fastest possible Internet service" is just plain false. Sure it may eventually be able to bring 1Gbps to homes, but fiber can deliver 100Gbps+ today. DOCSIS 3.0 or not, delivering data via electricity over copper has physical limitations, whereas sending data via light through glass has no limits. If we want to be able to enable space-age 3D experiences that go beyond just HD then we're going to need big broadband highways that can add lanes indefinitely to keep up with demand, and that means fiber.

"Also, much of the money devoted to wired broadband is meant for service that is at least 45 megabits per second, faster than all but a few homes can get now. But since the technology to deliver service that fast is already well developed, this may not have much impact other than making sure that when some people get broadband for the first time, it will be of the very fastest speeds."
Saul finishes his piece by glossing over one of the most important aspects of the debate over rural broadband. It's vitally important that as we're getting people online we're providing them with infrastructure that can deliver the fastest possible speeds both today and tomorrow. This may be our one shot for the next decade to subsidize broadband deployment to rural areas, so let's make sure we do this right the first time and bring broadband that's truly world class to every corner of our great nation.

---

In summation: Look, I'm not saying Saul's a bad guy who doesn't know what he's talking about and who may be on the cable companies payroll, he's just misguided. He's looking at broadband through the paradigm of today and not tomorrow. He understands the need to get big broadband connectivity, but wonders why spend all the money for fiber when you can get the functional equivalent in DOCSIS 3.0.

What he fails to understand, though, is that only fiber has the capacity to keep up with the demands of the 21st century. Only fiber delivers the reliability we need to created reliable networked experiences. And only by having fiber can we have competition in wireline broadband.

He also fails to touch on the upload side of the equation. Every DOCSIS 3.0 deployment that I've seen to date is still heavily asymmetric, with much slower speeds upstream than downstream. But if we continue going down this road we're going to enable consumers without empowering creators. We're going to encourage a digital economy where we only import rather than export digital content. We're going to miss the opportunity to encourage a generation of lean-forward mass media, where citizens actively engage in rather than passively watch the rebirth of America in the 21st century.

The simple, inescapable truth is that DOCSIS 3.0 or not, we can't build the next 100 years of our economy on a copper infrastructure; we need fiber. Anything less and we're committing ourselves to a future of being good but not great. And that's something I and others can't and won't accept.