Welcome to the Journal of Optical Communications and Networking (JOCN) OFC 2019 Special Issue. Published over two months (January and February 2020, Vol. 12, Nos. 1 and 2), the special issue includes extensions of invited and top-scored optical networking papers that were presented at OFC 2019. The OFC Technical Committee worked with the JOCN Editor-in-Chief to determine which authors were eligible to participate in this issue. The authors were required to include significant new material as compared to their OFC paper. All papers underwent the usual JOCN peer-review process. All of the authors, reviewers, editors, and JOCN staff are acknowledged for their diligence in producing this high-quality special issue.

As OFC is one of the premiere conferences for optical networking research, it is worth considering the drivers of this research. The most significant impetus comes from architectural advancements that have a profound effect on how networks are designed and operated. This was clearly demonstrated by the paradigm of optical bypass and transparent networks, a vision that had its beginnings in the late 1980s. This led to hundreds of optical networking papers that honed the architecture, proposed the requisite algorithms, and performed the associated economic analyses. A much more recent example is that of elastic optical networks, where a rigid fixed-resource approach is replaced by one with greater flexibility and adaptability.

A second important driver arises when the physical realities or limitations of the optical technology require that architectures and algorithms be developed in concert in order to attain high network efficiency. For example, the realization that optical bypass had its limits and that regeneration needs to be strategically deployed in continental-scale networks was essential to practical network deployments at the turn of the century. This resulted in a large body of networking research related to optimal selection of regeneration points. Another example is wavelength assignment algorithms that minimize the wasted bandwidth in a mixed line-rate network, necessitated by the potential detrimental interactions among the various modulation formats. More recently, the various incarnations of space division multiplexing (SDM), e.g., multi-fiber, multi-core, few-mode, have led to a need for associated algorithms, as well as new optical-switching architectures, to maximize the capacity benefits, address scalability issues, and optimally handle the wavebands that may be unavoidable due to coupling.

These two drivers of research certainly still exist, and will necessarily continue as optical networks evolve, although perhaps to a lesser degree. One field where the interplay between architecture and technology is still an activating topic is with respect to adaptable low-margin optical networks. In this paradigm the network elements are tuned to more precisely match the requirements of the current traffic, rather than employing fixed components or designing the network to meet worst-case scenarios. It is expected to deliver substantial network efficiencies; however, dealing with the practical challenges requires advancements in areas such as quality-of-transmission (QoT) estimation and distributed control to rapidly react to network conditions. Several papers in this special issue address some aspect of this paradigm, where attention has turned to techniques such as probabilistic constellation shaping (PCS) to more finely tune the performance parameters as opposed to the coarser tunability afforded by mQAM modulation. (JOCN published a special issue on low-margin optical networks in September and October of 2019.)

Currently, perhaps the most relevant network domain for the interplay between technology and architecture is the access network, most notably passive optical networks (PONs). Edge networking has become the new rallying cry across many industries, in part to deal with the range of exciting new applications that are envisioned (e.g., in 5G or 6G wireless). Roughly a quarter of the papers in this special issue address future PON directions. A major theme is how to simplify the technology and/or the architecture to meet the strict cost requirements of an access network while still meeting the stringent application requirements. (JOCN will publish a special issue on future PON technologies in Fall 2020.)

As has become evident at OFC, these historic drivers of optical networking research have been transcended by the drive to apply general networking techniques to the optical layer. The most obvious example of this is the introduction of artificial intelligence (AI) into optical networks. In this special issue, roughly one quarter of the papers propose using AI techniques, such as machine learning (ML) or deep reinforcement learning (DRL), to optimize optical network design and/or control. The applications range from QoT estimation to network orchestration. A second general networking paradigm that has made significant inroads into the optical layer is the softwarization/virtualization of the network. Another quarter of the papers in this special issue address topics such as software defined networking (SDN) and network disaggregation, sometimes in concert with AI, SDM, or PONs.

As you read the papers in this issue, please give thought to the drivers of exciting new optical networking research in the future.