1000Mbps Over Copper Telephone Lines: ITU Begins Work On G.fast, Successor to DSL
 |
| 1000Mbps Over Copper Telephone Lines |
The ITU has begun the process of approving G.fast, a new standard that will allow for access speeds of 1,000 megabits per second (1000Mbps, 1Gbps) over copper telephone lines. G.fast (full name ITU-T G.9701) is expected to be a cheaper and easier-to-deploy alternative than FTTH and FTTC (fiber to the home and cabinet). G.fast is the technology that will finally bring super-fast broadband speeds to the masses who are stuck with copper wires for the dreaded last mile.
There are generally three ways of getting internet access to your house: Telephone lines (twisted-pair copper wires), coaxial cable (shared with the cable TV infrastructure), and fiber (either to your house or fairly close). Because of the massive installed base of telephones and cable TV, the first two options are by far the cheapest. Fiber, because it generally has to be installed specifically for internet access, is much more expensive to deploy. The problem with telephone wire and cable internet access is that you can only squeeze so much data down a thin, low-grade piece of copper. This is where DSL, VDSL, and now G.fast enter the equation.
Back in 1999, the ITU standardized ADSL (G.DMT), a method of delivering up to 8Mbps down and 1.3Mbps up over standard telephone wires. To do this, ADSL used a special modulation technique that could utilize about 1MHz of audio bandwidth in the 25 to 1104 KHz range. Because the standard voice line only uses 30 Hz to 4 KHz, ADSL could be safely layered on top, allowing for voice and internet connectivity at the same time. Later, VDSL, ADSL2, and VDSL2 would use more advanced modulation techniques to use even more audio bandwidth. VDSL2, offering up to 200Mbps over 30MHz of bandwidth, is currently the best internet access that can be provided over telephone wire. In reality, due to physical and financial constraints, very few ISPs offer more than 100Mbps VDSL2 (17MHz).
G.fast hopes to raise the max link speed to 1Gbps, using 106 MHz of bandwidth. Again, due to real-world constraints, we’re probably talking about 500Mbps of combined (down and up) access speed. At these higher frequencies, cross-talk between the dozens of copper wires in a bundle will be an issue. Also, at the higher end of its spectrum usage, G.fast will interfere with FM radio frequencies (87.5 to 108 MHz). The ITU is working on a specification for G.fast equipment (ITU-T G.9700) so that such interference doesn’t occur.
The other problem with using such high frequencies is that, due to cross-talk, line noise, attenuation, and other physical constraints, the range of G.fast will be very short — on the order of 250 meters (820 feet). Basically, G.fast is envisaged as the ideal way of getting fiber-like speeds from the cabinet, manhole, or telephone pole — which in turn is connected to the ISP’s core network via a beefy multi-gigabit fiber link. By using the copper wires that already run to every home, the ISP is saved the massive expense of running fiber inside every home (which is estimated to make up 80% of the total cost of FTTH deployment). If you haven’t heard of it before, this is the “last mile” conundrum: It’s easy to run a big backbone to few cabinets, but running hundreds of separate wires/fibers into people’s homes is incredibly difficult.
The next step is finalization of the G.fast spec, which is expected to occur sometime in 2014, and then creation of G.fast-compatible chips and equipment, from the likes of Broadcom and Alcatel-Lucent. Actual deployments of G.fast could happen any time after that — but given how slow ISPs tend to be, you probably shouldn’t expect a 500Mbps internet connection until at least 2016 or later. As always, ISPs are trying to work out whether customers actually need 500Mbps. You can just imagine the thought processes of their CEOs: Surely those 100Mbps connections that we’re painstakingly rolling out will last you at least a decade, maybe two… right?
There are generally three ways of getting internet access to your house: Telephone lines (twisted-pair copper wires), coaxial cable (shared with the cable TV infrastructure), and fiber (either to your house or fairly close). Because of the massive installed base of telephones and cable TV, the first two options are by far the cheapest. Fiber, because it generally has to be installed specifically for internet access, is much more expensive to deploy. The problem with telephone wire and cable internet access is that you can only squeeze so much data down a thin, low-grade piece of copper. This is where DSL, VDSL, and now G.fast enter the equation.
G.fast hopes to raise the max link speed to 1Gbps, using 106 MHz of bandwidth. Again, due to real-world constraints, we’re probably talking about 500Mbps of combined (down and up) access speed. At these higher frequencies, cross-talk between the dozens of copper wires in a bundle will be an issue. Also, at the higher end of its spectrum usage, G.fast will interfere with FM radio frequencies (87.5 to 108 MHz). The ITU is working on a specification for G.fast equipment (ITU-T G.9700) so that such interference doesn’t occur.
The next step is finalization of the G.fast spec, which is expected to occur sometime in 2014, and then creation of G.fast-compatible chips and equipment, from the likes of Broadcom and Alcatel-Lucent. Actual deployments of G.fast could happen any time after that — but given how slow ISPs tend to be, you probably shouldn’t expect a 500Mbps internet connection until at least 2016 or later. As always, ISPs are trying to work out whether customers actually need 500Mbps. You can just imagine the thought processes of their CEOs: Surely those 100Mbps connections that we’re painstakingly rolling out will last you at least a decade, maybe two… right?
.jpg)
0 comments:
Post a Comment