We’re talking again today about the power of G-PON technology and the future of fiber networks. The highest bandwidth rate standard for passive optical networks is NG-PON2. It relies on time and wavelength division multiplexing which combines 4-8 wavelengths at 10 Gbits/s per wavelength downstream and either 2.5 Gbits/s or 10 Gbits/s of upstream bandwidth. NG-PON2 requires tunable filters and lasers at the ONTs. It can also coexist with each of the other previous standards. It allows for phased deployments by wavelength or multiple services to be segmented by wavelength.
Beyond NG-PON2 the IEEE has a task force working on 100G-EPON. It should provide for bandwidth speeds of 100 Gbits/s by increasing the per wavelength bandwidth rate to 25 Gbits/s. Currently the task force has an ambitious schedule to reach a standard in the next few years. There are a number of considerations that they are dealing with, backward compatibility and wavelength allocations. Allocation of wavelengths is starting to become a constraint as the standards all try to coexist in the limited spectrum available to maintain backward compatibility with the existing optical distribution network investments.
All of the work on these standards has been guided by the existing significant investments already made in passive optical networks. It seems that the FCC’s model for nationwide broadband and the history of optical distribution network investment will be directing the future of fiber optic technologies.
That is also supported by research being conducted by business and academia. Recently researchers were able to achieve transmission rates of 1 Tbits/s (terabits per second) over existing optical fibers. This involves using 11 light waves but only 1 receiver which resulted in no major changes needed to an existing optical network.
Another group of researchers achieved a record for the fastest optical data transmission over fiber. They achieved a rate of 1.125 Tbits/s by combining the 256QAM format used in cable modems with 15 different wavelengths of lights, each acting as a channel. This allowed the researchers to create a ‘super-channel’ enabling them to deliver that rate of data transmission.
Besides these amazing data transmission rates, researchers are also working on the limited spectrum that current standards are managing. A group of researchers have been able to use nanotechnology to create diamond Raman-Lasers on a photonic chip. Previously these types of lasers could not be created on integrated circuits making their use expensive. Diamond has much better optical properties than the silicon based lasers currently used across optical distribution networks. This potentially opens up new spectrum for use because the researchers were able to create the lowest operating power and longest wavelength produced in an on-chip Raman-Laser.
The research examples above show that efforts are being made to capitalize on the massive fiber optic distribution network investments that have been made. This makes logical and rational sense because the optical distribution network represents 70% of the sum of investments in passive optical network rollouts. The ODNs have a number of architectural factors that can alter that cost percentage. As mentioned previously the FCC is assuming G-PONs with distributed splitters in the field. A passive optical network designed that way locks a service provider in to always having a PON. Another option for passive optical networks is to have splitters located centrally. This creates a dedicated optical path from a node to each subscriber allowing the possibility in the future to put electronics in the node if there is a need to move to an active network. It also provides flexibility to alter the split ratios in a PON as the network grows. A third architecture is the central switch homerun which provides a dedicated optical path all the way back to the central office. The central switch homerun is the highest cost of capital architecture but provides the most flexibility in the future.
The investments being made in optical distribution networks are expected to have 25-30 year lifespans. Some of the optical distribution systems have already exceeded their lifespan projections. Bicsi and other professional organizations are advocating for 40 year useful life for outside plant. These generational investments mean that accurate record keeping and distributed documentation are vital for maximizing the investments made in these networks as management is handed over to future generations of employees. Those records will be vital for capacity planning and adjusting to future technology trends.
Lastly I came across some compelling use cases for fiber optic networks in rural areas in Europe. Farmers are using drones to photograph their acreage in multiple colors including near infrared wavelengths, these photos are then sent to the cloud for processing. As a result of the analysis farmers are able to increase yields in the near term of 15% per acre with expectations of long term increases of 50%. The significant data collection of these drones, some can collect as much as 4 GBs in a 10 minute flight require large bandwidth connections to the farm. Another use case is an automatic insect catcher called Scoutbox which takes 300 pictures a day, sends them to the cloud for counting and detection of harmful insects. Both of these use cases provide revenue or cost saving reasons to bring fiber to the farm.
The future of fiber optic networks looks to be progressing in incredible ways, and is a reflection of its past which should ensure that passive optical networks deliver the services that service providers will deliver into the future.