As is always the case with such work, development of CableLabs’ Coherent PON (CPON) specifications requires decisions on a number of elements. Curtis Knittle, who is overseeing the effort in his role as vice president – wired technologies at CableLabs, admits the decision-making process for CPON isn’t quite as far along as he hoped it would be when the effort began but says that hitting the original goal of a 2024 completion date remains possible.
CableLabs launched the CPON initiative last year with an eye toward supporting single-wavelength 100G over 80 km or with a split ratio as high as 1:512. “We had hoped to have our specifications done in the 2024 timeframe, and then have some interoperability events and be ready for deployment around 2025, 2026,” Knittle states. “We would need certain things to fall into place and then there's a possibility [to hit that 2024 goal]. But that was always an aggressive timeline when we started the project. So, it depends on how the cards fall in terms of some upcoming discussions that we're having with operators.”
Those discussions, both within the working group and between the group and the Operator Advisory Group supporting the effort, include how to split that 100 Gbps of capacity into smaller, more useful segments. Time and frequency division multiplexing (TFDM) is one approach currently on the table. “The Operator Advisory Group hasn't aligned on a specific mode of operation in that regard. So we're doing an assessment right now on use cases and the business requirements,” Knittle reports. “We're hopeful that will lead us and the working group in general to a decision.”
The groups also are evaluating options for the media access control (MAC) function. Knittle notes that other organizations, such as the ITU-T and IEEE, have developed MAC approaches for high-speed PON operation that CableLabs could leverage. “I'm reasonably confident that we don't want to define a different one – a third MAC. But that is an option,” Knittle notes. CableLabs is looking into joining the ITU-T’s specification development efforts on 50G PON, which may increase the chances that it won’t have to go its own way on the MAC.
Other elements that await consideration include the wavelength frequency at which the coherent PON will operate and whether full-duplex operation will be necessary. On the former, Knittle says that a wavelength frequency that enables compatibility with point-to-point approaches, such as the 100G and 200G coherent point-to-point schemes CableLabs has created (see “A Light Lunch: CableLabs Eyes Coherent Optics” and “CableLabs publishes coherent optics specs”), would be preferable.
Meanwhile, debate has not yet begun regarding whether to take a bidirectional approach such as the one CableLabs adopted for its coherent point-to-point specifications. Knittle says that work to validate full-duplex operation through a PON splitter has not occurred. Nevertheless, the benefits to a PON architecture should be the same as in point-to-point networks.
“The benefits of full-duplex coherent optics would translate whether it's point-to-point or point-to-multipoint,” Knittle explains. “One of the main driving factors [for full-duplex] is a single laser in each endpoint. One that's used as your local oscillator to demodulate the receive signal – and then that same wavelength, that same laser, being used to transmit in the opposite direction. And so when you think about a PON topology, other than the fact that the optical distribution network is passive and you have a power splitter in there, the same benefits arise.”
Fortunately for the specification work, interest in coherent PON is increasing, judging by the fact that five vendors have joined the CPON working group this year alone. Meanwhile, CableLabs is looking for engineers interested in assisting the organization in its explorations of coherent transmission.
It may not be possible to predict when network operators will require a coherent 100G PON, yet that day is coming, Knittle believes. “We do believe that the future in the access network is coherent,” he says. “At the rate capacity [demand] is increasing, as are the distances that cable operators need to transmit that optical signal, it's going be coherent.”
