G.fast is getting a large promotional push this month, and with good reason. The ability to deliver gigabit and faster speeds over existing copper is a powerful tool for bringing higher speed broadband faster and more cost-effectively than ever before. While people go goo-goo for G.Fast, let's not forget it will be fiber doing the heavy lifting.
The sweet spot for G.fast is between 100 to 300 meters, according to a presentation I saw at Broadband Vision earlier this month. After that, the current state-of-the-art for high-speed copper broadband goes down drastically as the length of wire grows. ADTRAN (News - Alert), Huawei, and a number of hardware vendors are hot on G.fast, with ADTRAN announcing G.fast trials this week with service providers in Europe, North America, and in the Asia-Pacific region.
Being able to deliver gigabit speeds over copper at distances of up to 300 meters is good for service providers because the vast majority of their commercial and home customers are within that distance, especially when it comes to non-U.S. carriers. Within that 300 meters might be a central office or a more humble distribution point in an above-ground box, but that 300 meter hard-stop point means carriers will need more fiber if they wish to deliver faster services to the largest number of customers.
AT&T (News - Alert) has to love G.fast. The company bet on fiber-to-the-neighborhood, with the last "foot" being copper, then set about building an architecture that relied less on trenching and more on short-run local loops to deliver broadband. G.fast provides AT&T and lots of European operators with a way to deliver better broadband without the expense of trenching fiber all the way to each and every house in an area.
But there are also plenty of companies that had basically written off copper—yes, I'm looking at you Verizon (News - Alert)—that are likely rethinking their strategies, especially given the encroachment of faster DOCSIS cable speeds. G.fast provides a bridge technology to deliver gigabit broadband speeds at a cost that more than cuts the difference between doing nothing and conducting a full-blown FiOS (News - Alert) installation.
The work for Verizon and others is to figure out the best strategy to deliver a "hybrid" network of fiber and copper into unserved and underserved markets, bringing in enough fiber and assorted network support to support gigabit copper delivery over existing infrastructure.
Since existing DSL deployments average anywhere from 500 Kbps to 15 Mbps download speeds and 384Kbps to 1 Mbps upload speeds, service providers planning to deploy to G.fast won't get a free lunch. Core fiber speeds to existing DSL edge deployments would increase by a rough factor of 100, with fiber moving closer into customer service areas.
I suspect a majority of G.fast deployments will generate both more fiber trenching—albeit a fraction of what you'd get if you went with a pure fiber-to-the-home deployment—and a lot of gnashing of teeth when it comes to core network upgrades. If the existing network within a service area is built to support multi-megabit asymmetric DSL speeds, G.fast deployment will require the backend to support symmetric gigabit services. That's fiber at anywhere from gigabit to 10 Gbps speeds and faster.
Edited by Maurice Nagle