You are going to start hearing about a new technology that may infuse some life back into existing copper networks. The technology is being referred to as G.Fast. This technology promises to be able to deliver very fast speeds up to a gigabit over copper for very short distances.
Some are referring to G.Fast as a last mile technology, but it is really a drop technology. The distances supported by the technology are so short that this is going to require fiber to the curb, or as some are now calling it, fiber to the distribution point.
Alcatel-Lucent and Telekom Austria just announced a field trial of G.Fast. That trial achieved a maximum speed of 1.1 Gbps over 70 meters and 800 Mbps over 100 meters for brand new copper. On older copper the speed dropped to 500 Mbps for 100 meters.
Current copper technologies use only a small portion of the theoretical bandwidth available on a copper wire. For example, most VDSL2 systems deployed today use up to 17 MHz of spectrum on the copper. G.Fast can provide more speeds by using more of the available spectrum and will be able to use somewhere between 70-140 MHz on copper. Plus G.Fast will be more efficient. Today DSL functions by dividing the data path into sub channels which each contain about 15 bits of data. Engineers are looking at coding and modulation techniques that will increase the bits per sub channel for G.Fast and thus increase speeds more.
G.Fast also will benefit by an existing technique called vectoring. This technology is used today with VDSL2 and eliminates crosstalk interference between copper pairs. It does this by monitoring the noise on copper and then creating an anti-noise signal which cancels the noise in the same was as is done by noise-canceling headphones.
Right now Alcatel-Lucent is spending a lot of time on G.Fast because they see a big opportunity to make more money out of the old copper networks. So let’s look at the issues that a large telco like AT&T will face when considering the technology:
- Because the distances to deploy G.Fast are so short, the carrier is going to have to build fiber past every customer, just like in a FTTH network. A large carrier like AT&T has some advantages over a fiber overbuilder in that they can overlash fiber onto existing copper on pole lines. This is cheaper and faster than putting up fiber for a new provider who has to deal with pole make-ready costs.
- Copper drops are generally the worst copper in the network. These wires get banged around by wind and suffer from repetitive water damage and are the weak point in the copper network. The promised savings from G.Fast is to lower the cost of installation at a customer. Some of this savings disappears if too many homes need a new drop to make it work.
- G.Fast will save the cost of getting into the house. Once connected to an existing telephone NID on the outside of the house the signal can go anywhere in the home that is already wired for telephone. But the distance issue quickly kicks in and I would expect carriers to take this to a wireless right inside the house.
- Savings are going to depend on how inexpensive the G.Fast electronics are compared to FTTH electronics.
- Large telcos have relied for years upon customer self-installation of DSL and they will need G.Fast to work the same way.
So the savings to somebody like AT&T come from a) cheaper fiber installation costs because of the ability to overlash, 2) the ability in many cases to use existing drop and inside telephone wires, and 3) the ability to have customers self-install the product to avoid having to go into the home.
There are still a lot of technical issues to consider and overcome. Some issues that come to my mind include things like overcoming existing splices in the copper, and making sure there is no interference with existing DSL.
The expected time line for the deployment of G.Fast is as follows:
- Standards finalized by spring of 2014.
- Chip sets developed in 2015.
- First generation hardware available in 2016 that probably won’t support vectoring.
- Mature second generation equipment available in 2017.
Since a carrier has to build fiber everywhere for this to work, the technology is really competing against FTTH. By the time this is readily available there may be lower-cost units for FTTH deployment and I think any carrier would prefer an all-fiber network if possible.