"Optical fiber is the only medium that will handle 10-Gigabit Ethernet."
"UTP has proven itself a sufficient infrastructure for the Ethernet standards developed to this point, right up to Gigabit Ethernet. But that's where it will end."
Who would make such incredible comments? (And I mean the literal interpretation of the word "incredible"—unbelievable; not to be taken seriously.) As it turns out, the person who made those comments is the same person whose face appears on this page.
Although the second quote above is paraphrased just a little, those statements appeared in this magazine's Editorial column in July 2001. I spent much of that column declaring, and then lamenting, the end of the copper-versus-fiber battles that waged between certain pockets of the industry (see "Why the copper-vs.-fiber war is over," July 2001, page 9). Apparently, two years is enough dormancy for those industry pockets, because I think I hear the rumble of battle stirring up again.
In the eight issues we have published so far in 2003, including this issue, we have devoted a healthy amount of space to the 10GBase-T efforts within the IEEE, the market's reaction to the possibilities of such a protocol, and the cabling-standards efforts that will be required. This month's contribution comes from Chris DiMinico, one of the copper-cabling devotees I was calling out two years ago.
The same day I wrote this column, I participated in a conference call with a market-research firm that cited 10GBase-T as having the potential to "take away a key driver for fiber deployment."
Earlier this year, we conducted an online survey of system end-users regarding 10-Gigabit Ethernet. In the survey, 43% of participants said their networks will require 10-gig speeds within five years. Another 31% said they will need that level of speed between five and ten years from now. Less than 8% said they need it right now, while 8.5% said they will need it more than a decade in the future, and 9.5% predicted they will never need that level of transmission speed.
We have all heard the expression "timing is everything." That could prove true when it comes to media choices for 10-Gigabit Ethernet. As much as we have discussed the topic lately, it is important to remember that the standard is still relatively early in the development stage. Some are predicting ratification in or around 2006. If you consider our survey this year to be a decent barometer, then at the same time the 10GBase-T standard becomes reality, a whole lot of system users will be making tough decisions about the infrastructure for their 10-Gig systems.
Oh, did I mention that 10-GbE is currently available in a fiber version? And while take-up is reported to be slow for now, any user upgrading to that speed in the next few years won't have a standardized twisted-pair copper option. See what I mean about the rumblings of battle stirring again?
When copper and fiber were duking it out in the go-go days of double-digit market growth in the late 1990s, it looked something like a science-fiction battle for world domination. Now, it is more akin to a post-apocalyptic squaring-off for the only remaining harvestable land. That's the reality check among all my high-flying discussion of high speeds and network upgrades.
But if you ask me (not that you did), I believe the battle is back on. I also believe the truism that good competition yields better options for the consumer. In that regard, I'm looking forward to it.