G.fast: From Promise to Gigabit Reality
“First, we must appreciate that a significant percentage of subscribers globally live in MDUs. Depending on the market, this typically ranges from 20% to over 70% of subscriber units in a serving area.”
In global markets, where a focus on speed has become an addiction as well as a necessity, traditional copper DSL technologies have been viewed as “outclassed” by newer, faster technologies like DOCSIS 3.0/3.1 and GPON. Even the most widely deployed copper technology today, VDSL2, in all its variations (that include vectoring, bonding, and a variety of profiles) can’t compete with these technologies, typically being deployed at a fraction of the speed of its competition. Then, just as DSL technologies seemed doomed to fade away, along came a wholly new copper technology that not only blew away traditional DSL performance, but allowed copper to catch up and even leap to the forefront of the speed game. Welcome, G.fast!
Since G.fast’s standardization in December 2014, the technology has rapidly moved from a promising technology over copper infrastructure to a viable and commercially deployable reality, providing a powerful new path for service providers who need to leverage their existing copper infrastructure.
Able to take advantage of a broader spectrum (up to 106 MHz), G.fast technology offers the promise of near-gigabit speeds over copper, and even higher with bonding. For service providers rich in outside copper plant or in markets with a high proportion of MDUs, it is truly a transformative technology. Like traditional DSL technology, G.fast is still subject to copper loop length, quality, and external noise interference. That said, the bandwidth upside of G.fast, both downstream and upstream, is worth careful evaluation.
Why is G.fast transformative? One of the main reasons is its speed. G.fast has delivered downstream speeds over 900 Mbps in loops below 70 meters in length in controlled labs and pilot network environments.1 Many commercial deployments have achieved similar speeds, especially those inside MDUs or apartment buildings.
Meeting MDU Demands
So why is G.fast such an attractive solution inside the MDUs for service providers? First, we must appreciate that a significant percentage of subscribers globally live in MDUs. Depending on the market, this typically ranges from 20% to over 70% of subscriber units in a serving area. For example, according to the latest U.S. census, there are nearly 30 million MDU family units in the U.S. alone.
Although many service providers have shifted their broadband deployment strategies in favor of fiber, it is not always possible, cost-effective, or logistically feasible to deploy fiber to each unit in an apartment building.
Universal access inside the MDU is also a major challenge for many service providers. Landlords and management companies are often reluctant to allow new cabling infrastructure work to disrupt MDU tenants, especially if there are already competitive alternative service providers within the building.
Added to that is the complexity of the seemingly endless variety of MDU types, which can make the time and cost to fiber an MDU hard to predict from one building to the next. Factors such as differing sizes, building materials, access to ducts and the riser (which makes the need for site surveys), and differing fiber deployment techniques, all need to be considered.
When combined, it is easy to see why MDUs can be a thorn in the side of many service providers, and often are underserved even by those that are the most forward thinking.
However, the ability to offer ultra-broadband services using the existing MDU copper infrastructure is understandably tantalizing for service providers, especially if solutions exist that can meet market demands for much longer than just the short term.
Commercial deployments of G.fast began to emerge in the spring of 2016, and have now reached enough of a critical mass to be both meaningful and educational.
A major lesson learned from early deployment is the need for solutions that can be deployed and managed easily. With the reduction in loop lengths necessary to reach desired G.fast speeds, service providers may see a five- to tenfold increase in the amount of managed network elements inside a service provider’s serving area. As a result, new network elements need to be fast to deploy, always on, and simple to manage and operate. It is anticipated that software-defined access solutions and SDN platform architecture will play a major role in helping service providers get this challenge under control.
Another major lesson is that MDUs offer a myriad of physical topologies and challenges, and that G.fast solutions need to be able to accommodate existing internal copper infrastructure, including existing CAT-3, coaxial cable, and even old CAT-1 station wire.
Although performance at the shortest distances does not differ greatly between all of these copper technologies, the inherent shielding of coaxial cable has meant service providers have witnessed very small drops in performance even at distances of up to 1,000 feet (300 M). In fact, over 800 Mbps downstream has been recognized at 1,000 feet over coax, while CAT-3 twisted pair performance has topped out at approximately half that speed at the same distance.
The third important lesson is that gigabit speeds can be achieved over copper when leveraging G.fast when bonding is utilized. For those operators seeking to offer “fiber-like” gigabit services, especially over existing CAT-3 telephone wires, the introduction of G.fast bonding has been a significant development. First deployments of bonded G.fast have demonstrated speeds of up to 1.5 Gbps, and, more importantly, assured services of 1 Gbps in the vast majority of MDU cases.
Service providers are finding that the exceptional bandwidth performance, flexibility across differing wiring types, and the ability to deploy inside the MDU as well as externally in environmentally hardened cabinets, are great advantages of G.fast. For example, the cost of an external hardened solution is proving to be a powerful option in solving building access problems or getting around the lack of an air-conditioned coms room or closet. (See Figure 1.)
Another valued feature service providers identify from G.fast is its ability to offer greater flexibility between upstream and downstream speed ratios than previous generations of copper technology or competitive DOCSIS solutions. Due to its inherent time division duplexing (TDD), Downstream (DS) and Upstream (US) signals use the same frequency to transmit across the wire. This means that service providers have the ability to choose what ratio of bandwidth they want to offer downstream versus upstream to subscribers, anywhere from 90:10 to 50:50 (US to DS).
What’s Next for G.fast?
The next major development for G.fast is just now arriving to market for commercial deployments: the introduction of Dynamic Timing Allocation (DTA). DTA introduces a major differentiator versus DOCSIS solutions because the ratio between DS and US allocated bandwidth over G.fast can change automatically and dynamically. What this means to the subscriber is symmetrical bandwidth on demand. Service providers can use DTA to dynamically deliver symmetrical speeds when necessary, then revert back to a downstream-biased service when not necessary. Clearly, this feature has many benefits for both the residential and business Cloud user alike.
2017 also represents an exciting time for G.fast in the DPU and FTTC environment. Two new standardized amendments are extending the performance rate and reach of the technology. This means that G.fast will expand its reach to deliver high-speed services at distances viable for existing cabinet infrastructures of over 2,000 feet from the home. These same enhancements will allow asymmetric performance on shorter loop lengths to deliver gigabit speeds in environments like MDUs even across a single copper pair.
Scalability options will also increase in 2017 with the introduction of 48 port G.fast board solutions and the capability of some solutions to locally vector up to 96 ports across the copper binder. These enhancements will be enormously beneficial in meeting the demands of larger high-rise MDUs and the needs of existing cabinets and DPU SFU requirements.
From promise to reality, G.fast has rejuvenated opportunities over copper, and enabled gigabit speeds to a whole new segment of the market. As we move through 2017, new enhancements will allow G.fast to move from niche markets into the mainstream.
1. Ovum/ NBN: nbn conducted its first G.fast trial in Melbourne last October and achieved total aggregate speeds of 600 Mbps over 100 meters of twisted-pair copper inside an MDU, having previously achieved speeds of 970 Mbps over 20 meters of copper on a lab trial in the nbn’s National Test Facility in Melbourne. http://www.nbnco.com.au/blog/industry/new-report-shows-glittering-future-for-g-fast.html)