The Challenges With Replicating G.fast Speeds in the Real-WorldDelia Hernandez
Internet Service Providers (ISPs) competing with cable service providers are watching the development of G.Fast, a new xDSL technology that promises to deliver fiber-like speeds using legacy copper lines.
G.Fast is the next generation of xDSL services that will benefit customer premises located near distribution points.
The increased usage of cloud-computing, video streaming, and the awaited arrival of the Internet of Things (IoT) revolution (expected to increase Big Data) have further increased the demand for faster broadband speeds. Internet Service Providers are seeking newer methods to improve broadband speeds to rates that exceed 100mbps to remain competitive. Internet Service Providers (ISPs) have been able to deploy high-speed broadband using legacy copper lines with xDSL technologies such as ADSL and VDSL. But with G.fast, ISPs are hoping to achieve near 1Gbps speeds in shorter loop lengths.
G.fast provides a wider frequency band than VDSL2. VDSL2 supports a frequency profile of up to 30 MHz. G. Fast’s frequency band currently supports 106 MHz profiles but newer generations of G.fast are expected to support 212 MHz frequency bands. These wider frequencies can provide near-gigabit speeds in laboratory settings.
G.fast Performance in the Real World
But the challenge for ISPs is replicating those laboratory results in real-world settings. Trial runs in real-world scenarios have successfully been able to achieve speeds of up to 330 Mbps.
UK’s Openreach Telecom company has deployed trial runs in Huntingdon, Cambridgeshire and received favorable customer feedback.
Huawei notes that one of the main issues with G.fast is minimizing the total cost of ownership. G.fast is still a bleeding edge technology and access service providers are experimenting with ways to maximize the commercial viability for G.fast.
G.fast’s higher frequencies overlap with VDSL and VDSL2’s broadband spectrum making it difficult for network architectures to support both xDSL access technologies.
Providers are faced with another challenge: G.fast support for short loop lengths entails drop point installations. Huawei notes that these types of deployments are characterized by their low-port counts.
Broadband providers are considering expanding port-count, but at the cost of sacrificing bandwidth. A higher port count will increase the amount of lines a DPU binder supports, and hence, increase the complexity of vectoring processes. According to Huawei, “the vectoring complexity of G.fast is 6 to 12 times [more complex than] that of VDSL2 Profile 14a with the same port count”. Nonetheless, the company remains optimistic concluding that “Over the next few years, vectoring processors that are more scalable are expected to become available.”
Huawei also anticipates that G.fast deployment will be particularly difficult in brownfields that boast a large amount of VDSL2 subscribers. ISPs will need to entice customers to upgrade to G.fast to be able to “re-farm… the spectrum to improve bandwidth after the VDSL2 CPEs are removed from the network”.
Getting subscribers to upgrade can prove particularly difficult if they find their existing VDSL2 speeds satisfactory.
Reverse Power Feeding
G.fast will rely on reverse power feeding to draw power from customer premise equipment to power G.fast DSLAMs. G.fast has even shown to require less power than VDSL2 with energy-efficient features. Reverse power-feeding will help reduce costs for distribution point installations.
TDD (Time Division Duplexing)
G.fast will rely on Time-Division Duplexing or TDD to transmit data. TDD uses pre-assigned time frames to alternate between sending and receiving signals. TDD gives ISPs the ability to customize the allocation of bandwidth between upload and download speeds and create asymmetrical profiles.
Take Our Poll!
If we were to roll out a new G.Fast DSLAM, how many ports would it ideally support?