Page 49: of Offshore Engineer Magazine (Nov/Dec 2024)

48 50

that it usually happens within a single data center. In the ENHANCED NETWORK MONITORING inference phase, in terms of network traffc, a pre-trained Coherent transponders and, to a lesser extent, the trans- model is queried using relatively short text prompts, and port platforms they are housed in offer a veritable fre hy- the responses sent back are also short text documents or drant fow of telemetry, with the latest transponders hav- still images. North-south traffc per user will obviously in- ing hundreds of information points in their published data crease as AI moves from text responses to still images and models. Dealing with this food of information over the then to ever-increasing quality of video. data communication network has always been a challenge, but this is exactly the kind of problem that AI can help

AI’S IMPACT ON SUBMARINE with by extracting correlations and patterns that the hu-

NETWORK TECHNOLOGY man eye alone cannot recognize.

Once a submarine cable is deployed, it has an engineer- ing life of 25 years, but the only thing we can upgrade PREDICTIVE MAINTENANCE during that lifetime is the submarine line-terminating Hand in hand with enhanced monitoring is the ability equipment (SLTE), and there may be multiple opportuni- to use the processed telemetry from the transponder, trans- ties to enhance cable performance by taking advantage of port platform, SLTE terminal, and repeater chain to detect the latest transponder and SLTE terminal technology. AI potential operational issues and to deal with them before can play a part in this enhancement in two different ways: they become a problem.

operational and network-based functionality.

NETWORK-BASED USE OF AI

OPERATIONAL USE OF AI Network-based features generally run in AI-enhanced

There are three main operational areas where AI can be transponders, and at least two different features have seen used to enhance SLTE capability, and all of these functions active development.

tend to be housed within the management and automa- tion software of the network. THREAT MONITORING AND

PROACTIVE PROTECTION

AUTOMATING NETWORK PLANNING For several years, Infnera has been involved in the devel-

AND CAPACITY OPTIMIZATION opment of techniques for seismic anomaly detection – ba-

In a modern open cable system, it’s essential to char- sically using existing submarine cables to enhance tsunami acterize each fber pair to ensure that the cable operator warning systems. Cable operators have shown interest in this got what they paid for in terms of wet plant performance. capability, but they have also thrown down the challenge of

Characterization using a range of parameters has been well enhancing sensitivity in order to detect threats such as sea- defned by the SubOptic Open Cable Working Group. bed trawling or dragging anchors in the vicinity of the cable.

But the process of taking these measurements can be labor intensive, and AI can certainly play its part in reducing the ADVANCED NONLINEAR COMPENSATION manual work required. The limiting impairments for modern coherent tran-

The next step for operators is spectrum planning and sponders are nonlinear effects caused by high optical allocation. That can include the need to support spectrum power levels in the fber. Using fully effective, algorithmic sharing on a given fber pair. This stage is performed know- nonlinear compensation (NLC) techniques is extremely ing the exact type of transponders that will be deployed, computationally intensive – well beyond the immediate and these performance parameters can be plugged into an roadmap for ASIC processing power. Limited NLC does

AI model that can help optimize the available spectrum. deliver useful additional performance, but one direction of

Finally, optical power planning is an ongoing aspect of research is looking at the use of a neural network approach fber pair operation, and it’s essential that this is carefully for NLC that uses far less processing power yet delivers a designed in order to ensure stable service levels. signifcant performance improvement.S november/december 2024 OFFSHORE ENGINEER 49