Wednesday, February 6, 2008

Stretching copper

When AT&T moved the goalposts back again on its expected deployment of bonded VDSL2, the industry seemed to collectively shake its head in frustration. Yet another delay in the carrier’s fiber-to-the-node initiative. And yet another return to the inconvenient truths about the bandwidth limitations of copper.

“AT&T will need to do something dramatic in order to find the bandwidth to deliver multi-stream high-definition [TV], especially since the cable and satellite competition have expanded their HD channel line-ups and/or VoD libraries substantially,” said Erik Keith, an analyst at Current Analysis. “MPEG-4 will help, but FTTN DSLAMs will not be the best strategy long-term. FTTP is really the best way to go.”

However, a few developments this week are aimed at not giving up on copper so quickly.

Rim Semiconductor is testing a new specification known as Internet Protocol Subscriber Line, or IPSL, that it expects will dramatically improve broadband performance over existing copper. Rim claims its Cupria processor can send traffic down a 26-gauge copper pair at 40 megabits per second over 5,500 feet. The company tested its wares a few months ago on the network of Monroe Telephone, a small telco in Monroe, Oregon not far from Rim’s headquarters in Portland.

John Dillard, Monroe’s president, told me Rim’s creation was “still in the breadbox stage,” but added that the Rim folks seemed to be very pleased with the results of their tests, promising more tests to follow.

Meanwhile, Alcatel-Lucent said this week that it expects to introduce equipment late next year that uses Dynamic Spectrum Management, or DSM, to aid VDSL2 performance by reducing noise in the network. But even that is described by Alcatel as a way to get a few more years of life out of existing copper before the inevitable transition to fiber.

These are only the latest such efforts to stretch copper’s limits. Last year, researchers at Penn State, working with cable vendor Nexans, designed a transmitter and receiver to send data at 100 Gb/s over Cat 7 copper.

Considering the expense of deploying FTTP and the amount of copper wire strewn across the country, it’s no surprise to see vendors focus on squeezing ever more bandwidth out of copper. As I’ve pointed out before, no matter what analysts say about the clear superiority of FTTP, the favorite triple-play access architecture of U.S. carriers continues to be FTTN–and copper the rest of the way.

No comments: