Mediterranean Shipping Company has overhauled its data architecture, moving global shipping operations from fragmented batch reports to near real‑time analytics. The Geneva‑based carrier is one of the first enterprises to adopt SQL Server 2025 in production, coupling it with Microsoft Fabric’s mirroring capability and database‑native AI to drive operational intelligence.

MSC’s IT leadership made the decision after years of wrestling with legacy systems that could only provide snapshots of vessel movements, container tracking, and port activities hours after the fact. In an industry where a single delay can cascade into millions of dollars in penalties, those stale views were no longer acceptable. The new platform, built on SQL Server 2025 and Fabric, processes telemetry from over 800 vessels and thousands of IoT sensors, delivering actionable insights within seconds.

SQL Server 2025 anchors the modernization

SQL Server 2025, currently in advanced preview, represents Microsoft’s biggest leap in on‑premises and hybrid database technology since the introduction of the intelligent query processor. MSC is running it on Azure Arc‑enabled infrastructure, allowing the company to manage SQL Server instances across its own data centers and multiple cloud regions from a single control plane.

The new version introduces a vector‑accelerated query engine, enhanced in‑memory column store with persistent memory support, and a built‑in REST API layer that makes real‑time data access feasible for containerized microservices. For MSC, the most immediate benefit came from the engine’s adaptive query processing improvements—queries that once took minutes to aggregate sensor readings now complete in sub‑second time.

“We needed a database that could keep pace with the volume and velocity of our IoT data,” said an MSC data platform architect, speaking on condition of anonymity because the implementation is still being optimized. “SQL Server 2025 gave us a 40‑fold improvement on certain time‑series aggregations without any schema redesign.”

Fabric Mirroring eliminates ETL latency

The second pillar of MSC’s modernization is Microsoft Fabric Mirroring. Fabric, a unified analytics platform, can now replicate entire SQL Server databases into OneLake with transactional consistency. Mirroring continuously synchronizes changes, making operational data available in Fabric’s analytical engine in near real time.

Previously, MSC’s data engineers had to build and maintain elaborate ETL pipelines that moved data from production SQL Server instances into a data lake for reporting. Those pipelines introduced latency of up to six hours and often failed during peak shipping periods. Mirroring voided that complexity. Once MSC enabled mirroring on their SQL Server 2025 databases, the data appeared in Fabric’s Lakehouse within seconds of commit. Analysts then built Power BI reports and Fabric Notebooks directly on top of the mirrored tables, confident they were working with up‑to‑the‑moment information.

“The elimination of ETL is not just convenience—it’s a fundamental change to our data supply chain,” the architect explained. “Vessel masters now see fuel consumption dashboards that reflect the last five minutes, not yesterday’s numbers.”

Database-native AI predicts disruptions

The most forward‑leaning component of MSC’s architecture is the use of database‑native AI. SQL Server 2025 embeds machine learning capabilities directly within the engine, allowing model training and inference to run on the same hardware that stores the data. MSC is leveraging this to predict port congestion, estimate maintenance needs for cargo‑handling equipment, and optimize route planning.

Model training happens inside SQL Server using T‑SQL’s enhanced PREDICT function, which now supports open‑source frameworks like ONNX and LightGBM without data leaving the database. For real‑time scoring, MSC has embedded inference calls into stored procedures that fire on sensor data insertion. When a refrigerated container’s temperature reading deviates from the expected pattern, the database immediately triggers an alert and adjusts the cooling unit’s setpoint without waiting for a centralized analytics job.

“Database‑native AI closes the loop between insight and action,” said a Microsoft spokesperson familiar with MSC’s deployment. “When milliseconds matter, moving data to a separate ML serving layer is too slow. SQL Server 2025 collapses that latency.”

Performance metrics and early wins

Since the phased rollout began in early 2025, MSC has reported measurable improvements:

  • Data freshness dropped from an average of 4.2 hours to under 2 seconds for vessel telemetry.
  • ETL pipeline failures were eliminated, saving the equivalent of 2.5 full‑time engineers.
  • Fuel consumption variance was reduced by 12% after database‑native AI models began recommending optimal engine settings in real time.
  • Report generation for port operations now completes in 300 ms, down from 47 seconds, enabling live dashboards in harbor control towers.

The system currently processes 5.2 billion row inserts per day across mirrored SQL Server 2025 primary replicas, with peak throughput of 120,000 events per second. Fabric’s OneLake automatically compresses and optimizes the data for analytical queries, and MSC’s data science team runs ad‑hoc explorations using Spark pools without impacting the operational workload.

Architecture that supports hybrid reality

MSC’s configuration reflects the hybrid nature of global shipping. SQL Server 2025 instances run both on‑premises at major ports and in Azure regions close to shipping lanes. Azure Arc manages policy, patching, and backup across these dispersed nodes. Fabric Mirroring links every instance to a central OneLake tenant, where a semantic model unifies the data for reporting.

This design allows each port to operate autonomously if connectivity to the cloud is lost—a critical requirement for maritime operations. The local SQL Server 2025 node continues to ingest sensor data, run AI predictions, and serve applications. When the connection is restored, mirroring catches up seamlessly.

Community reception and cautionary notes

Early adopters on the Fabric community forums have expressed excitement about mirroring but caution that the feature is still maturing. Some users report that large‑scale mirroring can incur significant compute costs if not properly tuned, and that the initial sync of multi‑terabyte databases can take hours. MSC’s team worked closely with Microsoft to fine‑tune the mirroring cadence and partition strategy, which mitigated most of these issues.

Security practitioners have also raised questions about how mirrored data inherits access controls. Microsoft’s documentation confirms that Fabric roles can be mapped to SQL Server logins, but granular row‑level security requires careful planning. MSC implemented a dedicated security layer using Fabric’s domain‑isolated workspaces to keep shipping‑lane data compartmentalized.

Feedback from the wider Windows and data community has been positive, with many pointing to the MSC case as proof that database‑native AI is ready for mission‑critical workloads. “This isn’t a research project—it’s a global supply chain running on SQL Server,” noted one Microsoft MVP on a popular Windows forum.

What this means for the broader ecosystem

MSC’s modernization is a blueprint for other asset‑intensive industries—logistics, manufacturing, energy—that rely on SQL Server but have been held back by batch‑oriented data pipelines. The combination of SQL Server 2025’s performance leap, Fabric’s zero‑ETL mirroring, and in‑database machine learning lowers the barrier to real‑time intelligence.

Microsoft is expected to announce general availability of SQL Server 2025 later this year, along with additional Fabric Mirroring connectors for PostgreSQL, Cosmos DB, and external clouds. For Windows shops that have invested heavily in .NET, SQL Server Reporting Services, and Active Directory, the new stack provides a bridge to modern analytics without a rip‑and‑replace migration.

MSC’s journey demonstrates that even the most traditional industries can transform when data moves at the speed of business. The shipping company that once relied on daily reports is now using the same technology to predict outcomes before they unfold—and that, in logistics, is the ultimate competitive advantage.