Why is this page text-only?

« Lafayette Becoming Most Wired Community in America | Main

July 27, 2009 9:52 AM

Logarithmic Bandwidth Goals For America's National Broadand Strategy

The debate surrounding what bandwidth goals should be set to guide the creation of America's broadband policies suffers from a gaping chasm between arguments for aspirational vs. "realistic" benchmarks.

Those in the "realistic" category often limit themselves to goals as low as 25Mbps, arguing that going any higher potentially rules out technologies like DSL, which may be easier/cheaper/faster to deploy.

On the other hand, those advocating for the most aspirational goals seem to have no end to the amount of bandwidth they think we need, with some calling for speeds beyond 1Gbps heading upwards of 100Gbps, which while seemingly way out there are actually the speeds being deployed today on our nation's fastest research networks.

Then alongside all this is the argument that says setting goals based on bandwidth reflects too much of an obsession on speed, that what really matters more than the capacity of a network is what it's being used for, and that goals should be based on usage. While I agree with this sentiment in general, I do think there's good reason to set bandwidth goals as the amount of capacity you have in large part determines what you're able to do with these networks.

Also worth noting is the argument that we need to take a more expansive approach towards setting goals that incorporates other metrics like latency and jitter. I absolutely agree with these notions, but this post is focused on tackling the challenge of setting goals related to bandwidth.

With this range of issues on the table, let's start by looking at some of the considerations that must be made in setting proper bandwidth goals:

Timeframe: New broadband networks should be thought of as infrastructure and therefore the long view needs to be taken about whether or not they're capable of keeping up with future demand. And yet government doesn't seem able or willing to think about the ramifications of its decisions beyond a few years into the future. Because of this, I think we should start with setting a goal for where we want to be in 2015. That's far enough in the future that it suggests we'll be using a very different Internet with very different bandwidth demands from the one we know today, and yet not so far into the future as to stretch beyond the horizon of policymakers' ability to consider them seriously.

Aspirational: A common refrain heard in these discussions is that we can never know what the future will bring, especially when it comes to the growth of the Internet. Taking this a step further, the Internet always seems to outgrow even the most ambitious projections. Ten years ago no one thought we'd be using as much bandwidth as we do today. So while we do need to be realistic in the goals we set, we should also not shirk from being aspirational as we always underestimate what the Internet will do.

Application-centric: We absolutely need to be able to justify any bandwidth goals we set with an explanation for why that much bandwidth will be needed otherwise this is all likely going to be considered as more of an academic exercise than a realistic attempt to quantify where our national broadband policies need to have us headed.

Globally Aware: If the point of having a national broadband strategy is to set America on a path to be globally competitive, to remain leaders in the development of the next generation of the Internet, then our goals must be informed by what other nations are doing, otherwise we may set ourselves up to be permanently trailing the rest of the world.

After taking all these factors into consideration, I think any discussion of what bandwidth goal we should be setting for America to achieve by 2015 has to start at 100Mbps.

The reasons for this stem from the categories listed above:

Application-aware: Earlier this year I proposed that we set the goal of having broadband capable of supporting a two-way HD video stream for every American, arguing that video applications are what's driving demand for bandwidth, that there's a host of uses for video from healthcare to education to business and beyond, and that we need networks robust enough to support lots of simultaneous usage. Turning back to the discussion of bandwidth goals, a true HD-quality video is encoded at 15-25Mbps today. More compressed video at lesser bitrates may claim to be HD but they're not, and uncompressed HD can get up into the Gbps. Therefore, a family of four needs at least 60Mbps and upwards of 100Mbps just so that everyone in the house can be using one HD video application at the same time. Plus there are a host of other simultaneous uses of broadband that put further demands on the network. So if you look at what the world will look like in 2015, where the use of HD video apps will be the norm, we need at least 100Mbps to every home to support the delivery of these applications.

(As a quick aside, some papers have done more in-depth research into the topic of future demands for bandwidth but have come up with puzzling results. Take ITIF's Need for Speed report, which laid out a realistic vision for how a home of the near future could be using next-generation connectivity, concluding that: "This home would easily consume more than 90Mbps of aggregate bandwidth [both directions]." And yet when it came to recommending optimal speeds for bandwidth goals, it set the bar much lower at 50Mbps down and 10Mbps up. Unfortunately this appears to be a case of prioritizing a "realistic" approach to what's possible in terms of broadband supply over acknowledging realistic analysis of how high broadband demand will grow.)

Globally Aware: While some countries like the UK are setting more "realistic" goals like universal 2Mbps service by 2012, those nations that are considered broadband leaders have set the bar at 100Mbps and are already beginning to go past that, like South Korea, which has charged itself with the goal of becoming a 1Gbps Nation by 2012. Because of this I'll be brutally honest in saying that if we set our goal at anything less than 100Mbps that I will consider that a national embarrassment. I just don't see how we can consider ourselves serious about being globally competitive with goals lower than 100Mbps, regardless of how unrealistic some may consider achieving them to be.

Aspirational: Now some might wonder why I'm not pushing for a goal of 1Gbps by 2015, but my reasoning is simple. Networks capable of delivering 100Mbps by 2015 are likely also going to be able to scale to 1Gbps relatively easily, so if we can achieve a 100Mbps Nation then we should be well on our way to achieving a 1Gbps Nation. Also in the conversations I've had with all but the most aspirational thinkers, support for setting aspirational goals seems to break down in between 100Mbps and 1Gbps. At this time, saying that every home will need 1Gbps within the next five years just seems to be too big of a step for most to consider. That's why I decided to set the aspirational bar at 100Mbps for our 2015 goal.

Now with a goal of becoming a 100Mbps Nation by 2015 in place, we can start working both backwards and forwards to set additional goals, which brings me to the concept of logarithmic bandwidth goals.

Let's start this discussion by simply considering what a logarithmic growth in bandwidth looks like:

1Mbps, 10Mbps, 100Mbps, 1Gbps, 10Gbps, 100Gbps

It's hard to ignore how well this lines up with the evolutionary trends of broadband.

To use most of today's Internet you need at least 1Mbps. (As another aside, I don't see how a decade into the 21st century we're still defining broadband based on Kbps. Talk about embarrassing!)

To use all of today's Internet you need 10Mbps, including watching "HD" video on sites like Hulu.com, which requires 7Mbps.

And to be able to fully participate in the Internet of just a few years from now, where interactive HD video applications are everywhere, households will need at least 100Mbps.

Then looking out further ahead, speeds of 1Gbps, 10Gbps, and 100Gbps line up perfectly with where discussions about future bandwidth goals should be set.

So all that seems to be left is to associate this logarithmic growth in broadband capacity with the years we should set as goals for their widespread availability in the US.

For 1Mbps, we really need to get that everywhere ASAP. In fact, if we really had our act together, I'd argue that having 1Mbps be universally available should have been a goal set for 2006 or even earlier.

For 10Mbps, I'd like to be aggressive and say 2010 but some may see that as being unrealistic given that wireless technologies aren't yet delivering those speeds in the US. If that's the case then we should look at having this benchmark be set for 2011 or 2012 as an intermediary step.

For 100Mbps, I didn't pull the 2015 goal out of my hat as striving to become a 100Mbps Nation by then has already been proposed by Sen. Rockefeller. And the more I think about it the more sense it makes. It gives us an aspirational goal that we can realistically achieve in the next 5-6 years if we set our minds to it.

After that, I'm hesitant to set any additional goals. For one, policymakers will have a hard time seriously considering them. For another, it's somewhat foolish to think we can really guess where things are going to grow to that far into the future. So I'd rather not start making guesses now that will almost certainly underestimate what the future will bring.

Because of this, instead of trying to set out goals that are further into the future, I'd just commit ourselves to readdressing our bandwidth goals sometime in the next few years.

The final point I want to make about this issue of bandwidth goals is the debate over how we should quantify and set goals for coverage. All too often we get stuck in discussions over whether setting goals related to 100% availability are realistic given that even basic telephone service isn't necessarily universally available. But the flip side of this is that if we start setting goals like having 100Mbps available to 90 or 95% of the population, then we're basically saying we're OK with leaving millions of Americans behind. Plus there's also the twist that some of these goals may suggest wasting money bringing these networks to people who don't want them and won't use them.

So instead of getting lost in this debate, I'd suggest we instead simply focus our attention on making these speeds available to all Americans who want them.

The most important thing to remember throughout all of this is that if we want America to continue to be a great nation, we must be setting goals that put us on a path towards greatness. We can't allow efforts to be "realistic" to get in the way of striving towards aspirational goals. And we can't forget that with enough will, unified effort, and careful planning, any goal we set can be realistically achieved.

As I've said before and will continue to say again and again: We are America, gosh darn it. We can achieve anything we put our minds to. The only thing limiting us is our own imagination and self interests.

So let's acknowledge the clear trends of bandwidth-hungry technologies and our aspiring global competitors to set the goal of becoming a 100Mbps Nation by 2015. Because the sooner we're able to do that, the sooner we can start getting down to the real work of figuring out how to achieve this goal.

Del.icio.us Digg Yahoo! My Web Seed Newsvine reddit Technorati


TrackBack URL for this entry:

Comments (3)

Your opinion piece seems implicitly to accept the premise that government shouldn't be picking technology winners and losers. So the only way you can lobby for fiber-to-the-premises (FTTP) technology is to lobby for a bandwidth goal (100 Mb/s symmetric) high enough that you think only FTTP can deliver it.

In the case of first-mile telecom infrastructure, I think it's appropriate for government, in its National Broadband Plan, to pick FTTP as the technology we want, based on its ability to support not only 100 Mb/s symmetric speeds now but also a series of exponential speed improvements over the fiber infrastructure's useful life.

More specifically, I think the National Broadband Plan should pick point-to-point fiber infrastructure as the infrastructure we want. (Both active Ethernet and PON technologies can use point-to-point fiber infrastructure.)

It may be debatable whether the point-to-point fiber infrastructure should be one-fiber or two-fiber. I'm leaning toward two-fiber. IEEE has standardized point-to-point two-fiber optical transceiver interfaces (PMDs) for 100 Mb/s symmetric, 1 Gb/s symmetric, and 10 Gb/s symmetric; and PMDs for 40 Gb/s and 100 Gb/s should be standardized by next year. IEEE has also standardized point-to-point one-fiber PMDs for 100 Mb/s symmetric and 1 Gb/s symmetric; but I'm not aware of plans to standardize higher speeds.

For point-to-multipoint technology, where bandwidth is shared among a number of users, I'd propose that for the purposes of the National Broadband Plan, the bandwidth be said to be shared equally among the users. So, for example, if 32 users share the bandwidth of a 2.4/1.2-Gb/s GPON link, each user gets 75/37.5 Mb/s, and that's how fast the National Broadband Plan should say it is; note that this is less than 100 Mb/s symmetric.

If government wants to have a plan for wireless, let them come up with a National Narrowband Plan.

Posted by Jeff Hoel on July 27, 2009 6:08 PM

Jeff - Everyone knows I'm a pro-fiber guy, but I actually set about constructing this argument without fiber in mind. I tried to clearly state why we need to have our goals start at 100Mbps without getting into what technologies can deliver it.

If I'd been doing nothing more than creating a pro-fiber argument I would've set the bar at at least 1Gbps symmetrical as at 100Mbps there are other broadband technologies that can at least claim to deliver those kinds of speeds.

But I wanted to be ideology free when it came to technology in this discussion, basing the 100Mbps goal on the use of these networks and what our global competitors are doing.

That said, you are spot on that once we strip away the sacred cow of trying to hold onto technology neutrality from broadband policy debates and take a critical eye towards the many clear technological trends that point to a future that will require abundant bandwidth, I do think the case for a full fiber future becomes crystal clear if we want to be serious about being globally competitive in the 21st century digital economy.

Posted by Geoff Daily on July 27, 2009 6:18 PM

Affordable 100 Mbps. You could probably get 100 Mbps almost anywhere now. You just can't afford it.

Posted by Susan Estrada on July 28, 2009 2:18 PM

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)