Opinion: Beginnings of a bubble?

This article originally appeared on fibresystems.org.

Bob Metcalfe, often called the “father of Ethernet”, will be delivering a keynote speech at the OFC/NFOEC show tomorrow about the inevitability of Terabit Ethernet, which he says will get started by 2015.

And before that today, George Gilder will be giving a the closing keynote presentation at the 2008 Executive Forum, which is organized by the Optical Society of America (OSA) in conjunction with media house Pennwell. Although the title has not been revealed, you can bet your bottom dollar it’s going to have something to do with the “exaflood” — the collective wave of data that’s going to overwhelm the existing Internet infrastructure in the near future unless there’s significant investment in new equipment.

Isn’t it exciting to see such big numbers being bandied about? An “exabyte” is one thousand petabytes, which in turn is one thousand terabytes, or put another way, it’s 1018 bytes of data. To put that into perspective, that’s roughly 50 thousand times the amount of information in the US Library of Congress if it were held in digital form.

It turns out that Gilder and Metcalfe have a history of predicting runaway data traffic growth and the ensuing Internet doomsday. Writing in an InfoWorld column in 1995, Metcalfe calculated that the Internet backbones carried a mere 15 terabytes of traffic per month, while an estimated 15 exabytes of data might be circulating in the world’s local area networks. If those exabytes of data found their way onto the Internet, there would be a “catastrophic collapse” the following year, he predicted.

(Just for fun, it’s worth recalling that in 1997 Metcalfe admitted he’d been wrong. At an April conference, he literally ate his words, putting a copy of his article in a blender and drinking the resulting pulp.)

Metcalfe’s use of the term exabyte led Gilder to coin the term “exaflood” in 2001. And now its Gilder’s turn to sound the alarm bells. In an article co-written with Brett Swanson of the Discovery Institute published last month, he said that the Internet is about to drown in an exaflood of data created by new applications and services like movie downloading, IPTV, ubiquitous mobile cameras, online gaming, virtual worlds, and telepresence.

Both of these men, who seem to be the most vocal fortune tellers of massive Internet growth, are fronting the industry’s premier annual optical networking event. Now, I don’t know about you, but hearing these prophecies of spiralling data growth gives me a definite sense of déjà vu. Just seeing Gilder’s name on the programme takes me straight back to the hype of 1998 (when, incidentally, I’d just landed my first real job as a reporter with FibreSystems Europe).

Gilder is a self-confessed believer in the “build it and they will come” theory of telecoms. Similarly, on an interview with Light Reading TV, when asked about why we might need Terabit Ethernet, Metcalfe simply said “We build it, they will come, I’m sure. It’s happened every time for 35 years.”

Excuse me, but wasn’t it exactly this attitude that got the industry into trouble in the late 1990s?

The network operators built it and the Internet users did come… eventually. But for many companies and their investors the wait was too long, and they faced bankruptcy before their ideas could make any money. Much of the intellectual property survived somehow, but a lot of shareholders, which in many cases turned out to be pension funds, lost a lot of money. What’s to stop this scenario from playing out again?

Some may argue that this time it is different because the demand is more immediate. Internet service providers say they are suffering because they can’t build out their networks as quickly as they need to. On the other hand, “this time it’s different” could turn out be the most expensive words in the English language.

Obviously the Internet is going to grow, but getting the predictions spot on is a bit like predicting the weather — you know it will stop raining in San Diego at some point, but will it be on Monday like the weatherman says? Oh hang on, its stopped already (smile).

So, is this time any different? Is the telecoms industry going to repeat the mistakes of the bubble years or has it learned its lesson? Let us know your views by using the commenting tool at the bottom of this article.

Reproduced with permission. © Institute of Physics and IOP Publishing Ltd.