We are big believers in taking fiber all the way to the customer, said Lee Hicks, Verizon, speaking at ADTRAN's Broadband Business Solutions event in Huntsville, Alabama. Verizon's One Fiber program is its multiyear strategy to build a common fiber plant for all services, including consumer, business and mobile backhaul. The carrier starts with a deep base fiber, having invested in its FiOS FTTH network for years.
In April 2017, Verizon awarded a contract with Corning to purchase up to 20 million km (12.4 million miles) of optical fiber each year from 2018 through to 2020, with a minimum purchase commitment of $1.05 billion. At the time, Verizon cited an initial deployment of its One Fiber approach in Boston, where it plans to invest $300 million over six years to deploy it throughout the city.
Hicks said NG-PON2 is the right platform going forward for all services, in part because of its ability to support 4 x 10G wavelengths. NG-PON2 is described as a fraternal twin of XGS-PON because it uses the same electronics but more advanced tunable optics. Hicks said Verizon is seeing good results in a field trial of NG-PON2 in Tampa, Florida. There is a somewhat difficult upgrade path from Verizon's installed base of GPON consumer connections, but NG-PON2 could be deployed as an overlay in installed areas. Hicks said NG-PON2's tunable technology and link budget physics have been solved. Vendors need to address the engineering challenge of making NG-PON2 cost effective. ADTRAN is pursuing this opportunity. The next step will be vendor qualification tests.
Verizon will leverage this physical plant upgrade cycle to simplify its aggregation architecture. Circuit emulation will be used to convert legacy services to Ethernet. The carrier forecasts a 90% reduction in its router footprint.
The new Multiservice edge (MSE) promises a significantly lower cost per transported bit. Verizon will also eliminate the many probes it currently uses for service assurance. All elements of the new network will collect streaming telemetry data natively and store it in a common data lake.
Hicks noted that when Verizon first proposed its new architecture, there were plenty of skeptics inside the company. Now the question has become "why can't you go faster in implementing the changes?"
In April 2017, Verizon awarded a contract with Corning to purchase up to 20 million km (12.4 million miles) of optical fiber each year from 2018 through to 2020, with a minimum purchase commitment of $1.05 billion. At the time, Verizon cited an initial deployment of its One Fiber approach in Boston, where it plans to invest $300 million over six years to deploy it throughout the city.
Hicks said NG-PON2 is the right platform going forward for all services, in part because of its ability to support 4 x 10G wavelengths. NG-PON2 is described as a fraternal twin of XGS-PON because it uses the same electronics but more advanced tunable optics. Hicks said Verizon is seeing good results in a field trial of NG-PON2 in Tampa, Florida. There is a somewhat difficult upgrade path from Verizon's installed base of GPON consumer connections, but NG-PON2 could be deployed as an overlay in installed areas. Hicks said NG-PON2's tunable technology and link budget physics have been solved. Vendors need to address the engineering challenge of making NG-PON2 cost effective. ADTRAN is pursuing this opportunity. The next step will be vendor qualification tests.Verizon will leverage this physical plant upgrade cycle to simplify its aggregation architecture. Circuit emulation will be used to convert legacy services to Ethernet. The carrier forecasts a 90% reduction in its router footprint.
The new Multiservice edge (MSE) promises a significantly lower cost per transported bit. Verizon will also eliminate the many probes it currently uses for service assurance. All elements of the new network will collect streaming telemetry data natively and store it in a common data lake.
Hicks noted that when Verizon first proposed its new architecture, there were plenty of skeptics inside the company. Now the question has become "why can't you go faster in implementing the changes?"
