Want your network to be relevant in 2026? Perhaps it's time to look to the past for guidance. In June of 1997, Raj Reddy of Carnegie Mellon University issued a white paper called A Gigabit National Data/Fiber-to-the-Home Initiative for Computer Mediated Communications.
Having won the 1994 Turing Award, the highest honor in computer science, Professor Reddy was already known as a bona fide visionary. Reading his white paper, which quite closely envisions what has come to pass in the deployment of gigabit FTTH networks, only reinforces Reddy's reputation.
In his paper, Reddy put forth the argument that the US should build a national data grid with a fiber-to-the-home access network as the ideal solution to relieve the significant strain on the fragile voice telecom infrastructure created by the growth of Internet traffic; an endeavor that "appears to be an attractive business proposition, as well as important initiative to ensure the nation's leadership."
His estimates on the cost to deploy the technology, developed with Paul Shumate at Bellcore, are accurate today. Further, he projects that gigabit residential services will be available to consumers for well under $100/month and maybe as low as $50/month -- figures that bookend the range of prices for residential gigabit services today.
The white paper is most impressive in its prophetic accuracy of what will happen to companies that don't jump on the fiber future. He writes: "Most of the telecom and media giants will become the Western Unions of the 21st century, unless they are nimble and jump on the new bandwagon."
The specific standard that would start to fulfill Raj Reddy and Paul Shumate's vision and cost estimates we know today as Gigabit Passive Optical Networks (GPON), developed by the Full Service Access Network (FSAN) working group and standardized by the International Telecommunication Union, Standardization Sector (ITU-T) in the years following Professor Reddy's white paper. Since being introduced in 2003, GPON has been deployed to hundreds of millions of homes worldwide.
While GPON will continue to dominate FTTH deployments through the end of the decade, the technology that will bring a gigabit to most of the world's consumers -- NG-PON2 -- is just now being readied for commercial deployments. Why NG-PON2 is the future of FTTH (and not XGS-PON or 10G EPON) lies at the heart of Reddy's vision -- multi-wavelength optical networking.
While single-wavelength PON standards XGS-PON and 10G EPON enable higher bandwidth than today's GPON networks, they have none of the new service-enabling features and operational advantages that make NG-PON2 a service provider's ideal destination and choice for the next decade of optical fiber access.
Here are a few examples of how NG-PON2, with ONU wavelength mobility, will change optical access networks as we know it.
Capacity management through wavelength assignment
Because NG-PON2 ONUs utilize tunable optics, PON capacity can be easily managed and redistributed across initially four -- and, in the future, eight -- NG-PON2 wavelengths. Fast switching times between wavelengths means that subscribers can be moved from one wavelength to another without service impact. A boon to long-term capacity planning is evident, but it also brings into existence new time-of-day services and load balancing. For the first time, physical resources within the access network will be capable of being modified to meet the dynamic needs of the subscriber; and all without human intervention or a truck roll.
Service separation through wavelength independence
Because business and residential services have different SLA expectations and traffic behavior, service providers have been hesitant to share single wavelength PONs for these different customer types. While having more bandwidth in XGS-PON and 10G EPON helps, it doesn't solve the entire problem. NG-PON2 allows residential and business services to share a common optical infrastructure, but avoids contention through separation on different wavelengths. Operations and equipment isolation can also be tailored to individual wavelengths by having different OLT line cards delivering each service.
Equipment protection using wavelength switching
Because of the rapid ability to move from one wavelength to another, ONUs are able to sense when there has been an equipment failure at the OLT and can switch to a different OLT laser wavelength. As a result, full protection switching will be available on NG-PON2 systems, meaning higher reliability. Service providers that had not used PON systems for business services due to reliability and failover concerns will now be able to implement true carrier class solutions. Taking it a step further, two OLT systems in two serving offices could coordinate ONU wavelength assignments to enable full equipment and fiber path protection for high priority businesses and high-density multi-unit buildings.
Wavelength channel bonding
XGS-PON and 10G EPON are both limited to 10G speeds -- less than 9 Gbit/s when you consider the mandatory Forward Error Correction (FEC) function stipulated in the XGS-PON standard. Multi-wavelength NG-PON2 will have the ability to bond multiple wavelengths to achieve up to 40G and eventually 80G of bandwidth to a given end point. This expands the applications available to PON networks and when coupled with high reliability protection switching, opens up a much larger and more lucrative subscriber base for PON-based networks. The specifics of wavelength channel bonding are still being worked on by FSAN, but the technology is being demonstrated today.
Open network by wavelength
Open access networks around the world have always struggled with the best approach to creating a level playing field, while at the same time sharing a common infrastructure. NG-PON2's multiple wavelengths mean that service providers will no longer need to share bandwidth on a PON -- the network operator can instead allocate each service provider their own wavelength.
In single-provider networks, this same capability can be used to better leverage infrastructure and deploy equipment from multiple suppliers. Service providers will no longer be limited to one OLT/ONU solution in an optical serving area. Instead, they can deploy multiple systems, with each equipment vendor being assigned one or more wavelengths. This will free service providers to quickly deploy new features available from one of their equipment suppliers, without regard to the existing solutions in a given geography.
Bringing fiber-based gigabit service to every subscriber was visionary in 1997. Today, it is table stakes in competitive markets, and widely available in many rural areas thanks to service providers that saw the potential of fiber to foster an environment for social and business entrepreneurship. Professor Reddy's remarkable vision is well on its way to being realized, but there is still much more to do on a local, national and global scale. The complete vision spelled out in the 1990s for multi-wavelength FTTH were fulfilled by FSAN in creating the NG-PON2 standard. From the perspective of service providers, the advantages of NG-PON2 are compelling.
The optical access network of the future will be essentially unconstrained and the opportunities limitless. NG-PON2 is the technology to get us there.
— Alan DiCicco, Solutions Marketing Senior Director, Calix Inc. (NYSE: CALX)