By Kourosh Amiri, VP Marketing, Ikanos, Inc.
Demand for high-speed broadband access by consumers has never been more intense than it is today. Rapidly increasing numbers of connected devices inside the home and the adoption of higher-resolution (4K and 8K) television are just the tip of the iceberg. Home automation, remote patient monitoring, and multi-player gaming – among countless other applications – are contributing to an Internet of Things phenomenon that promises to drive bandwidth demands through the roof.
And carriers and ISPs are lining up to get a share of the prize, wondering how their current broadband technologies can evolve to meet the increasing demand. In the case of telcos, for example, even with the potential of vectored VDSL2 to deliver aggregate bandwidths of up to 300 Mbps to consumers, competitive pressure continues to mount to deliver quantum increases in bandwidth. To that end, G.fast, with its promise of up to 1Gbps service to each household, could roll out in initial trials in as little as 12-18 months.
G.fast, a concept proposed in 2012 and achieving consent by the ITU-T standard body in December 2013, represents the next performance node in the evolution of xDSL. G.fast is defined to support up to 1Gbps over short (i.e., less than 100 meter) copper loops, and is designed to address gigabit broadband connectivity on hybrid fiber-copper carrier networks. Service deployments are targeted from fiber-fed equipment located at a distribution point, such as a telephone pole, a pedestal, or inside an MDU (Multi-Dwelling Unit), serving customers on drop wires that span a distance of up to 100 meters.
G.fast – in the same way as existing ADSL and VDSL – takes advantage of fiber already deployed to cabinets or other nodes. The proliferation of G.fast will come as service providers push fiber closer to homes, where a single distribution point will serve, typically, from 8 to 16 homes.
Why the interest in G.fast? Even with vectored VDSL2’s ability to deliver hundreds of Mbps for Fiber to the Node (FTTN) applications (150Mbps aggregate performance on a 500-meter loop length demonstrated in many carrier lab trials worldwide by Ikanos, and 300Mbps for shorter loops), the explosion in the number of devices per home and new services and applications is expected to drive strong interest in accelerating FTTdp with G.fast to market. And, for carriers, the need to spur adoption of G.fast by their subscribers will play a lead role, keeping them in lockstep with competing services over cable and FTTH in the race to 1Gbps residential broadband connectivity.
Fortunately, much of the work in preparing the G.fast standard has already been completed, and chip suppliers have the consent of the ITU-T to start the development of G.fast chips, with some already making public announcements about their upcoming products. (For example, Ikanos in October 2013 announced an architecture and development platform for G.fast.)
The infrastructure for G.fast is a variant of the infrastructure for VDSL2. The primary difference is in the length of the copper pair that enters the residence. G.fast will require shorter copper loops than those used with VDSL2 in order to accommodate the desired gigabit performance. That, in turn, means that service providers must drive fiber closer to homes. In addition, carriers will need to ensure that the fiber-to-G.fast media converters at each distribution point will be backward-compatible with VDSL2. Why? To enable customers not yet subscribing to G.fast service (when it becomes available) to continue to receive VDSL2 service through G.fast transceivers in the network. This is a critical requirement for carriers looking to offer new services to their existing subscriber base. And it is a practical consideration, as not all subscribers will choose to upgrade to these new services at the same time. Without this VDSL2 backward-compatibility feature (also known as VDSL2 fallback), this transition to new services may end up creating many problems for carriers, including additional CAPEX and service disruptions.
As the adoption of G.fast nears, customer pre-qualification and self-installation are needed to ensure a smooth, cost-effective migration to G.fast. Existing customers may be able to leverage new ADSL2 or VDSL2 CPE (or a software upgrade to their existing CPE) with advanced diagnostics to pre-qualify the line for G.fast service. Those diagnostics would be used to qualify the copper pair, check for RF noise, and advise whether any line conditioning actions would be required when installing the G.fast DPU. What’s more, carriers expect that G.fast deployments will take place with virtually “no touch” installation and provisioning, setting the stage for more rapid adoption of the technology.
With the G.fast standard expected to be fully ratified later this year, and while deployment costs are still a matter of conjecture, and will potentially vary by geography, carrier, and specific network topologies, expectations are that they will undercut the costs of FTTH significantly. And, with the current uncertainties in service providers’ plans and timing for broad rollouts of FTTH, the future for G.fast looks bright.
Kourosh Amiri has more than 20 years of experience in the semiconductor industry. He has been responsible for the successful introduction of products targeting a range of applications in the networking, communications, and consumer segments. He joined Ikanos on February 2013 to lead the company’s global marketing and product strategy.Amiri was previously with Cavium, where he led marketing and business development for its emerging media processor group, and drove the strategy for turning Cavium into a leading supplier of wireless display media processors in multiple market segments, including smartphones and PC accessories. Prior to joining Cavium, Amiri held senior marketing and business development roles at Freescale and several venture-backed semiconductor start-ups, addressing a wide range of networking and media processing applications. Amiri has an MSEE from Stanford University and a BSEE from the University of California, Santa Barbara.
.
