Copper Wire As Fast As Fiber?

Krishna Dagli writes to tell us that a new consortium of hardware vendors and phone companies have banded together in order to try for fiber optic speeds over copper wiring. From the article: "To avoid interference, current DSL implementations use static spectrum management that is built for a 'worst-case' scenario. Most actual phone lines would allow for far better performance, and DSM technology will allow each DSL connection to be regulated in real time by the hardware based on measured crosstalk and on current data needs of each customer. The end result could be DSL connections that top out at 100Mbps or more."

Comments

Re:Sailing effect

[moderators_are_w*nke]

This story has been annoying me all day. Fibre does indeed have a much greater upper limit, so this whole story is complete fud. The following is a direct quote from Professor Andrew Tanenbaum (and he knows stuff) from his book 'Computer Networks' (3rd Edition, Prentice-Hall 1996):

"With current fibre technilogy, the achievable bandwidth is in excess of 50, 000 Gbps (50Tbps) and many people are looking very hard for better materials. The current practical limit of about 1Gbps is due to our inability to convert between electrical and optical signals any faster."

This was written in 1996. We've come a long way since then. Copper is simply not in the game.

Electric signal propagation speed in copper

[tygt]

From http://www.copper.org/copperhome/HomePlan/puffs_sm oke_pulse_electrons.html, which is the copper trade group and hopefully reasonbly accurate and not overly optimistic, I find:

Such waves would travel at the speed of light except that they are slowed down slightly by the effects of the insulating material surrounding the wire. Speeds of one-third to more than one-half the speed of light are typical.

OTOH, http://www.itarchitect.com/article/NMG20010416S000 6 states:

In more ordinary media, such as certain commercial single-mode optical fiber products, the propagation velocity of a signal is 68 percent of c or 205,000km/s ... In comparison, electric waves or signals in commonly used copper wire travel at speeds between 55 percent and 80 percent of c.

So don't take it for granted that just because an electric signal doesn't travel at c in copper that it's slower than light in fiber!

On a barely-related tangent: As someone who put up with a satellite internet connection for 4 years, I can state authoratatively that the speed of light isn't nearly quick enough for a variety of purposes....

Re:Light faster than Current, but does it matter?

[thebdj]

One thing I will add to the parent post that should be considered is noise and interference. Electrical signals are more susceptible to noise and interference. This is probably the biggest advantage fiber has over electrical transmission. I actually believe many of the other "advatanges" of fiber are in some way linked to this advantage.

Re:*scratches head*

[MoralHazard]

"Fast" can refer to either bandwidth (how much data the pipe carries at one time), or latency (the delay between when a signal is sent and when it is received). Here, they're talking about bandwidth.

But as other have pointed out, the summary is QUITE misleading. Copper wire of a given spec (length, guage, etc.) has a maximum theoretical bandwidth. The actual bandwidth we get out of it depends on the aim and sophistication of the signaling mechanism. My 3.0 MBps DSL line uses the exact same copper pair of phone wires as a dial-up 56k modem would, but it achieves much greater bandwidth because it crams more signal onto the wires. For last-mile distances (which may be more or less than a mile, obviously), the current practical maximum is much higher than current DSL rates, but we would have to use a more sophisticated signaling mechanism than DSL does.

Optical fiber has similar theoretical limits, governed by length and cable spec, and similar practical limits depending on how the signaling works. If we're talking about realistic cable guages and last-mile distances, optical fiber will always have a higher theoretical bandwidth than copper wire--that's just physics.

HOWEVER... since the practical maximum is governed by the signaling technologies in use, the theoretical maximums don't really matter.

Performance isn't the issue here.

[m.dillon]

You can shove more data over fiber, by several orders of magnitude, given enough equipment at the end-points. Copper only has a few hundred megaherz worth of useable frequency spectrum and that coupled with the noise floor and signal to noise ratio (SNR) puts an upper limit on how much data you can push over it. You can't just pump in higher voltages to improve the SNR because even with variable-sized twisted pair you will get noise leaking into adjacent wires.

The issue with fiber is that getting the several orders of magnitude improvement in bandwidth requires increasingly expensive equipment at the end-points. This is fine for long-haul fiber but obviously not appropriate for a consumer end-point. Fiber gets multiplicative bandwidth improvements by transmitting light at different frequencies all over the same physical fiber optic cable. Specialized chips can pick-off the frequencies and split them into individual transceivers. A consumer end-point could decode one of those frequencies fairly cheaply, but not much more then that before the equipment becomes expensive. This is certainly viable... the head-end can transmit dozens of frequencies over the fiber and the distribution point on the pole can split it out to homes, or even just route it over shorter-haul copper in proximity to the home (which is probably a lot cheaper then running fiber all the way into a home).

-Matt

That would be very cool.

[Kadin2048]

Actually in some other discussions I've said that I think this could be really beneficial -- with systems like BitTorrent, and to a lesser extent Skype and other P2P systems, it's conceivable that a big broadband provider could configure its network so that a lot of "bulk" traffic was kept on its own wires, and didn't have to traverse the public net.

For example, if they provided a Skype supernode that all the broadband users could connect to, whenever one of the customers wanted to call another, the routing could all be done without having to send packets through a peering/transit point. It would all be on the ISPs network, which costs them basically nothing.

You can make similar arguments for positioning cache servers for other types of stuff on the network. Were it not for the copyright concerns, they could probably save themselves a lot of customer aggravation and bandwidth expense, if they just did some intelligent caching of bittorrent traffic. (And it's my understanding this is the whole theory behind the Cache Discovery Protocol, but I'm not sure which ISPs are going to use it.)

The place where I think this could have the biggest effects, would be in places that have large networks that are basically isolated from the public net by narrow connections -- say, Australia. A system of intelligent caching and encouraging the use of P2P applications would probably lighten the load on the traffic actually passing in and out of a "network island" by favoring internal connections instead.

So a broadband ISP that let you connect to your neighbor at 100Mb/s but only pass packets out to the public 'net at 1Mb, might at some point in the future, if it was designed correctly, seem like a really sweet deal.