NEC and Google apply AI technology to subsea cable networks

21 May 2018

Submarine cable being installed. Image: Vismar UK/Shutterstock

Japan’s NEC and Google have completed transmission tests across a subsea network using AI.

The ‘Faster’ trans-Pacific submarine cable system connects Japan to major US coastal hubs, from Los Angeles to Seattle and Portland. Six international companies, including China Telecom Global and Global Transit, are involved in the project, which has been overseen by Japanese firm NEC.

Faster by name, faster by nature

The cable, completed in 2016, is 11,000km long and its connectivity extends beyond Japan to other Asian locations.

This month, joint research from NEC and Google found that the spectral efficiency across the Faster cable system can be upgraded to 6 bits per second per hertz, for a capacity of more than 26Tbps. This is more than two-and-a-half times the originally planned capacity.

Google and NEC collaboration

Google and NEC said the tests used near-Shannon probabilistic shaping at a 64 quadrature amplitude moderation (QAM) modulation in combination with AI.

NEC added: “For the first time on a live cable, artificial intelligence was used to analyse data for the purpose of nonlinearity compensation (NLC). NEC developed an NLC algorithm based on data-driven deep neural networks to accurately and efficiently estimate the signal nonlinearity.”

“In doing so, the authors set a spectral efficiency-distance product record of 66,102b/s/Hz in a field trial performed together with live traffic neighbouring channels.”

Toru Kawauchi, general manager of NEC’s submarine network division, said the approach uses machine-learning algorithms that can be used on any submarine cable system. “The results demonstrate both an improvement in transmission performance and a reduction in implementation complexity,” Kawauchi said.

He gave further details about the potential for AI to boost connectivity in submarine cable systems: “Furthermore, since the black-box model is built up from live transmission data, it does not require advance knowledge of the cable parameters.

“This allows the model to be used on any cable without prior modelling or characterisation, which shows the potential application of AI technology to open subsea cable systems, on which terminal equipment from multiple vendors may be readily installed.”

The results of this field trial were presented at the post-deadline session of the Optical Fibre Communication Conference and Exhibition in San Diego, California.

Telecoms firms are racing to deploy subsea cable infrastructure as the global demand for data continues to grow.

Ellen Tannam was a journalist with Silicon Republic, covering all manner of business and tech subjects