The Coming ExaFlood

How Much is "Enough?"

Bret Swanson of the Discovery Institute gave this nascent debate the name "ExaFlood" in his January 2007 Wall Street Journal editorial, entitled The Coming ExaFlood.

The word exabyte is the basis for the term "ExaFlood", the neologism created by Swanson. The Exaflood refers to the rapidly increasing torrent of data transmitted over the Internet - it's a classic case of demand for bandwidth out pacing supply. The amount of information people upload, download and share on the Internet is growing at an exponential rate (due in large part to video, audio and photo applications). Meanwhile the capacity of the Internet, its bandwidth, remains limited and susceptible to a "flood" of data.

At the Broadband Properties Summit 2007, a panel on the ExaFlood raised a challenge.

The panel: David Kozischek, Corning Cable Systems; and David McClure - U.S. Internet Industry Association; Larry Irving, Internet Innovation Alliance.

The Challenge: How will the Internet handle the estimated traffic in the near to mid-term?

The key point driven home in this fascinating keynote panel discussion was this:

Not one, but several signs indicate a growing issue, if not a coming crisis.

Whether the ExaFlood represents an an issue, a crisis, or a market opportunity is a matter of perspective, however. Whether there is a growing need for bandwidth is not.

The FTTH Council Video on the ExaFlood, summarized in the Signposts inset above and shown below, provides a quick overview of the themes of this discussion.

A Matter of Scope

When faced with any problem, it helps to put it into perspective before spending too much time on a solution. When it comes to the "ExaFlood," how much do we know about data and the internet?

The amount of traffic on the Web has grown dramatically over its short lifetime, and it continues to show strong signs of growth.

Growth is driven by three primary factors: first, number of users - how many access the Web daily; second, frequency - the number of times each user accesses the Web; and third, bandwidth, the type of content that flows over the Web.

All three of those factors are on the rise, but it's the third - the type of content - that is most worrisome. The rise of video content, especially high definition video, dramatically increases the amount of data traffic.

According to Doug McClure of the US Internet Industry Association, print is not the problem, per se. Where a typewritten page is about 2,000 bytes, and a small graphic image is about 100 kilobytes, the complete works of Shakespeare about 5 megabytes, and an entire pickup full of books about a gigabyte, it would take a billion of those pickup trucks to amount to a single exabyte.

It's video that is the bandwidth hog. Standard definition video consumes 450 times the bandwidth of a regular web browsing session, but high definition video bumps that to 2,700 times the bandwidth.

YouTube Growth

JANUARY 29, 2007 YouTube videos comprise two percent of all Internet traffic. Surveys: Internet Traffic Touched by YouTube

JUNE 19, 2007 In six months, YouTube video traffic has grown from two to ten percent of all internet traffic. YouTube Comprises 10% Of All Internet Traffic | WebProNews

FEBRUARY 1, 2008 Six months later, YouTube video traffic comprises 20% of all internet traffic. Alexa.com YouTube traffic statistics

However one looks at it, this is a statistically significant growth curve that supports the concerns about the ultimate capacity of the internet, fast becoming as critical an infrastructure as our electric grid.

The Consequences of Growth

As more and more Internet traffic is video traffic, and as more and more subscribers start using broadband to gain access to the Internet, we will face some mixture of these consequences, to an ever growing degree.

First, without significant investment in infrastructure, continued growth in network traffic is likely to result in network performance degradation, most likely in the form of slower uploading and downloading.

One significant consequence is pressure to manage the Demand Side (i.e., Net Neutrality debate). As one follows this debate, one quickly recognizes the political agenda - there usually is a political agenda when looming shortages and the need for a change in public policy enter the picture.

The ExaFlood and Net Neutrality debates are - inevitably - interwoven, and predictably, political. A driver of the conditions that create the Net Neutrality issue in the first place is the scarcity of broadband bandwidth. Incumbents argue for the need to differentiate on service pricing in the face of scarce bandwidth. The ExaFlood is all about scarce infrastructure.

Moves by ISPs to suspend heavy users or cut them off entirely reflect the sense of scarcity. Political pressure to allow ISPs to charge varying rates for different levels of service or on the other side, to prohibit different levels of service are each signs of looming issues.

Incentives to manage the Supply Side may be a preferred alternative. While there will be increasing business incentives to build more last-mile broadband infrastructure, political incentives may be another way to stimulate broadband infrastructure growth that would relieve such scarcity.

Clearly, when it comes to broadband, efforts to fit all this increasing traffic into an already tight pipe (i.e., current network capacity) must be contrasted with some effort to figure out a way to get a robust sustainable infrastructure built out.

Anyway this issue is sliced, we're going to have to spend more capital on broadband infrastructure. We should be talking about the best ways to get this infrastructure built.

Fiber v. Wireless Broadband

In the face of ever growing demand for bandwidth, we need both the infinite capacity of fiber broadband, and the ubiquitous mobility and utility of wireless broadband ... we need to meet the ever growing need for both bandwidth and mobility.

We need wired solutions like Cable, and DSL, and wherever possible, FTTH, for reliable connectivity when we are in a fixed position, like at home or at work.

We need wireless solutions like cellular wireless and wireless broadband when we are mobile and away from our normal fixed locations.

Hot Spots, by the way, are not really adequate substitutes for a mobile solution. They're a portable solution - not mobile, but an alternate fixed location while roaming.

Wireless and wired broadband should be considered two equally valid, complementary technologies. Asking which is better pre-supposes that there is a certain need being filled and that wired and wireless are comparable solutions to fill that need. Most times, one is better than the other - the variable is the need for capacity v. the need for mobility.

In considering the need for both FTTH and wireless broadband - not one or the other - an apt analogy is " The 747 and the Helicopter."

The 747 is hard to beat if the objective is to move a large number of people economically and comfortably over a very long distance, say, over an ocean or between continents.

The helicopter is the best solution if the objective is to move a small number of people or things from any one location to another, generally over a relatively short distance, irrespective of airports and runways.

There is no better network for capacity than fiber - FTTH offers nearly infinite capacity and long-term viability because of its physics. But there is no better network for mobility than wireless broadband - Wi Fi Mesh offers a relatively low-cost wireless solution that goes up relatively quickly, then provides far more bandwidth than a cellular air card solution to far more people.

Which should come first, capacity or mobility?

On the one hand, a small town may opt for a FTTH solution first, because they have very poor alternatives for triple play services, and little need for mobility. But after the network is in place they may grow more interested in a Wi Fi Mesh network, because with fiber widely available, the price of Wi Fi Mesh goes down, and its utility goes up.

On the other hand, another town, perhaps a little larger than the previous one, may opt for wireless broadband because they already enjoy adequate competition for triple play service, but their city government operations could benefit from more options for mobile solutions. But after the wireless mobile broadband solution is in place, the value that the community places on broadband goes up, and there may be a more compelling appreciation for the benefits of FTTH.

It's a matter of priorities and at some point, after some experience is gained, it's about awareness and perspective. But in every case, we must come to some agreement that we all will need both the infinite capacity of FTTH and the ubiquitous mobility of wireless broadband, before all is said and done. Some of us will get it sooner than others. And those local populations that adopt broadband lifestyles, whether fixed or mobile, are more likely to devote the resources needed to get the other.

The Debate on Solutions

Advocates of Net Neutrality tend to argue in favor of flat pricing for access and separation of infrastructure ownership and/or management service from content delivery service (structural unbundling). Opponents of Net Neutrality argue for continuation of status quo vertical integration, variable pricing options to ensure quality of service and network management flexibility and little-to-no regulation to ensure adequate (abundant) revenue sufficient to finance network construction. There is significant distrust on both sides of the argument, and lots of history.

Source: Wikipedia article on "Bit"

But back to the approaching shortage of bandwidth. The first question is whether one believes that there even is a coming shortage. Debates about other shortages seem more long-term - the coming oil shortage as we approach Hubbert's Peak, the theory about approaching finite limits of oil production on the planet; or projections of world population growth and coming food shortages and famine (see also the Limits to Growth by the Club of Rome and Thomas Malthus).

The difference is that while the Club of Rome's projections of disaster and calamity have been proven wrong (so far), and Hubbert's Peak remains an unknown (so far), the ExaFlood dilemma seems more immediate and real, and we can still imagine that the problem is somewhat manageable.

Drowning in the ExaFlood gives an interesting counterpoint to this suggestion of a coming "flood" of data transmission shortage. Author Tim Lee, a fellow at the Libertarian Cato Institute, makes the argument that we are facing less of a "flood" of bandwidth shortage and more of a simple extension of the continuing need for telecom firms to build out their infrastructure, which supports his rationale for government to get out of the way.

Bring On The Exaflood by Bruce Mehlman, former assistant secretary of commerce under President Bush, and Larry Irving, former assistant secretary of commerce under President Bill Clinton - both currently co-chairmen of the Internet Innovation Alliance - embrace the problem and see solutions at hand, as follows.

All sides agree that we need ongoing investment in content, massive upgrades of infrastructure and relentless innovation to handle the phenomenal growth in data traffic.

We need advancements in how we build and operate networks, including new file compression technologies, upgraded traffic management software, better spam and virus filters, and new delivery platforms. And we need substantial investments in short-haul bandwidth through fiber to homes, broadband over power lines, satellites and fourth-generation wireless networks.

The formula for encouraging such extraordinary investments is clear: minimize tax and regulatory constraints and maximize competition.

Policymakers across the nation have ample opportunity to implement this blueprint right away. They should pass common-sense legislation such as permanently extending the Internet tax moratorium, building broadband-ready public housing, and cutting depreciation schedules for network equipment and infrastructure.

The Bottom Line

The ever growing demand for bandwidth appears to be quite real, but also incremental, so it need not be a crisis. But it deserves much greater attention and more creativity.

The ExaFlood scenario makes a strong argument on all sides to build out real broadband infrastructure at both the backbone (transport) and last mile levels (100 Mbs - 1Gbs) ASAP. Such speeds mean fiber in all its forms - nothing else will get us those speeds and be truly future-proof, or in this case, Exa-Ready.

And the need for infrastructure means letting go of limiting the number of players in the infrastructure business. We need more energy and creativity to accomplish such a Herculean task.

Conferences and articles may be good to raise the hue and cry and get people motivated and busy. But this is a project no less exciting than building the railroad network was in the 19th Century, and it demands the attention of leaders at all levels.

We should be building out broadband networks, both wired and wireless, as robustly as possible. Given these trends of exploding data needs, most especially the growing popularity of video applications in all their forms, we appear to be at little risk of overbuilding (certainly out on the edge we are not). We are quite likely to see an unending need for bandwidth in the foreseeable future.

ATTRIBUTION REQUIREMENT - The reader is encouraged to link to this post and to share this material, with the understanding that they provide an attribution to MetroNetIQ and share any alterations with MetroNetIQ.

This work is licensed under a Creative Commons Attribution-Share Alike 3.0 United States License.