João Pedro, an IT researcher and Senior Principal Engineer at Infinera, and André Souza, PhD student at IST and Senior Engineer at Infinera, won the Best Paper Award at the 14th edition of the International Workshop on Resilient Networks Design and Modelling (RNDM 2024). The workshop, which took place from 25 to 27 November in Pompeii, Italy, was organized by the University of Salerno, Italy, in cooperation with Gdansk University of Technology, Poland.
The winning study “Impact of channel provisioning strategies in the transient resiliency of SuperC+L-band Networks” describes a framework to model the interplay between design, operation, and failure simulation in a wideband optical network. Using this framework, it evaluates how key decisions made at the design stage and when provisioning services over the network will impact service availability. It offers insight on the main trade-offs between service resilience to transients and overall network cost/capacity metrics.
The increasing need for capacity in optical fiber networks, expected to be further accelerated by the impact of massive AI workloads on inter-data center traffic requirements, will inevitably require more fiber spectrum. Exploiting additional and wider transmission bands would enable to leverage of the existing fiber infrastructure and postpone the capital-intensive rollout of new fibers. However, in such wideband optical systems the impact of simulated Raman scattering (SRS), which consists of power transfer from higher to lower frequencies, is very pronounced. Besides the additional complexity of designing the network taking SRS into account, the availability of services running over these networks can be compromised due to SRS-induced power transients in the event of failures, even if such services do not traverse the failed network segment(s).
RNDM is an annual event focusing on network resiliency/failure survivability covering a wide range of related topics, from modeling, simulation, performance evaluation, and business impact to novel architectures, design/provisioning methodologies, and optimization algorithms.
More: