Splash247: The ship does not need to upload everything it knows

Published by Splash247

As vessel data volumes explode, the argument is shifting from how much information can be sent ashore to where intelligence should actually reside. The emerging answer is neither edge nor cloud, but knowing which decisions belong where.

For years, maritime digitalisation followed a simple assumption: collect data onboard, send as much as possible ashore and let the heavy computing happen in the cloud. That model is starting to look unnecessarily blunt.

Modern vessels can generate enormous streams of machinery, navigation, performance and environmental data. Faster satellite connectivity makes transmitting more of it possible, but that does not automatically make it sensible. The more important question is increasingly whether the information needs to leave the ship at all.

Tore Morten Olsen, president maritime at Marlink, describes edge computing as one of shipping’s most underestimated technologies.

“By processing, securing and prioritising data at the edge, operators can make faster decisions, optimise network performance and build greater operational resilience,” he tells SplashTech.

The obvious candidates are decisions where milliseconds, seconds or independence from the satellite link matter: machinery anomalies, cyber events, navigational inputs and other safety-critical functions. An edge server can ingest thousands of signals, identify the abnormality and send an alert or diagnosis ashore rather than continuously transmitting every raw reading.

Send the insight, not the noise

Gert-Jan Panken, general manager and vice-president at Inmarsat Maritime, argues that maritime needs to stop equating more data transmission with greater digital maturity.

“The objective should not be to act on every data point but to identify which data can materially improve a decision,” he says.

Across shipping, Panken believes the percentage of vessel data actually influencing operations remains relatively small. The challenge is making selected information reliable, timely and contextualised.

That logic lends itself naturally to edge processing. High-volume sensor streams can be cleaned, prioritised and analysed onboard, with the vessel transmitting the exception, trend or recommendation rather than terabytes of undifferentiated information. Panken’s preferred investment is the infrastructure capable of managing data intelligently between ship and shore: multi-network connectivity, edge capabilities, cybersecurity and orchestration.

The approach also makes a vessel less dependent on perfect connectivity. Offshore operations cannot assume the network will always be available, regardless of how dramatically satellite bandwidth has improved.

Where cloud wins

The cloud has different strengths. Fleet-wide analysis requires seeing across ships. Long-term machinery degradation, benchmarking sister vessels, comparing fuel performance, identifying recurring equipment failures and training models against years of operating history become more valuable when information from multiple assets is combined.

StormGeo’s Petter Andersen stresses that useful operational intelligence comes from combining different datasets — manual reports, sensor information, weather, performance data and vessel models — rather than treating individual streams in isolation.

The cloud consequently becomes the place for aggregation, comparison and learning; the edge becomes the place for filtering, immediate analysis and operational continuity.

That boundary will not remain fixed. AI models may increasingly be trained ashore but deployed onboard, while selected results return to headquarters to improve the fleet-wide model.

Not edge versus cloud

Olsen argues that presenting the issue as a contest misses the point.

“Cloud versus edge isn’t as important as the ability to secure intelligent orchestration across both,” he says.

That may be the more useful architecture for shipping. Process locally when speed, resilience, bandwidth or safety demands it. Send information ashore when aggregation across ships creates additional intelligence. Store what has future analytical value. Discard what is merely noise.

The digitally mature vessel will therefore not be the ship transmitting the most data. It will be the one that knows what needs to leave.