Microsoft Fabric and Azure data platforms in Dendermonde: architecture, capacity and cost

Before deciding on a lakehouse, it is worth establishing whether your data actually outgrew an import model or simply outgrew the person maintaining it.

This page is about building a data platform for an organisation in Dendermonde, Microsoft Fabric, Azure Synapse, Data Factory and the lakehouse patterns underneath them, with the local economy taken seriously. Dendermonde's base in logistics and freight, food and drink manufacturing, and chemicals and process industry determines the data volumes, the latency requirements and the retention obligations, and those three things decide the architecture. The client population here is shaped by logistics on the Scheldt and rail corridors and food and chemicals processing.

Does a Dendermonde organisation actually need Fabric?

The case for Fabric gets strong when more than one of these holds: import models under strain, the same data needed through both SQL and notebooks, governed self-service for several teams, retention that a desktop model cannot carry, or a Premium capacity already on the bill. One analyst maintaining three import models that refresh in twenty minutes does not need a capacity, and buying one will not improve anything.

What the data actually looks like in Dendermonde

Food and drink manufacturing

Short shelf life, high SKU churn, promotional volatility and retailer scorecards that measure you on their definitions rather than yours. Source systems typically include ERP, MES, warehouse management, retailer EDI and scorecards and quality systems.

For a platform design, the consequence is specific. Internal service levels read green while the retailer scorecard reads amber, because the two measure different events. So the architecture has to make the raw grain replayable rather than aggregating at ingestion, and the first serving model should be a production and waste model aligned to retailer scorecard definitions, so commercial conversations use the customer's own numbers.

Construction and real estate

Project cost and progress data with subcontractor valuations, variations and retentions, plus asset and tenancy data on the property side. Project reporting runs monthly on manually consolidated spreadsheets, so a problem contract is identified two months after it turned.

Systems to plan for: project costing systems, ERP, BIM and design systems, site progress apps and property management systems. Subcontractor payment and tax reporting obligations apply, and health and safety records carry their own retention duties.

Chemicals and process industry

Continuous process data at high frequency, batch genealogy, yield and energy intensity, with safety and environmental reporting drawing on the same sources. Historian data is retained for months while investigations and environmental questions reach back years. REACH, CLP and environmental permit reporting all require auditable activity data with a documented calculation method.

A medallion layout for a food and drink manufacturing business

The medallion pattern is not complicated, and it is routinely implemented badly: a bronze layer that has already lost fidelity, a silver layer with no testable rules, and a gold layer that is really just the silver layer with different names. Here is how we would lay it out for a Dendermonde client in food and drink manufacturing.

Capacity sizing and what it really costs

Fabric capacity is billed by capacity unit. An F2 is roughly US$0.36 per hour pay-as-you-go, about US$260 a month, with around 40 per cent off for a one-year reservation. An F64, the smallest SKU at which report consumers no longer need an individual Pro licence, is roughly US$8,400 a month pay-as-you-go or around US$5,000 reserved.

The arithmetic that matters: F64 only becomes cheaper than per-user licensing somewhere in the region of 350 to 600 readers, depending on the discount you actually pay and whether you reserve capacity. Below that, Pro or PPU per user is normally the right answer, and buying capacity for prestige is a straightforward waste of money.

Capacity is sized from measurement, not from a spreadsheet. Each workload is profiled for its compute shape, for background refresh against interactive query, for concurrency and for when the peaks actually fall, then the result is checked on a trial capacity running your own data. The output is a sizing model that states expected utilisation at launch, at double the readers, and the specific point at which you would need the next SKU. Always price from the current Microsoft rate card for your region and agreement. Any figure in an article older than a quarter should be treated as an order of magnitude, not a quote.

Direct Lake, import or DirectQuery

Direct Lake removes both the import refresh and the DirectQuery hop by reading Delta in place. It also carries a condition most teams discover late: unsupported constructs silently fall back, and the model then performs like DirectQuery. We test it against an import model on your own data and choose on the measured result. In practice Direct Lake tends to win on large, frequently landing tables, which food and drink manufacturing workloads often are, while import still wins on smaller models with heavy DAX.

Migrating from what you have now

Few Dendermonde organisations start from nothing. The realistic options are a serving-layer-first adoption, where Fabric takes the gold layer and reporting while existing engineering stays in place, or a domain-by-domain migration with the old platform retained until each domain is validated. Both avoid a big-bang cutover, which is the pattern that produces the rollbacks we get called in to repair. The reference pipeline we build first is usually the system that hurts most, which here is normally ERP.

Residency, latency and the Belgium question

Most Belgian deployments land in West Europe, which is physically in the Netherlands, or in France Central. Microsoft has announced a Belgian datacentre region, so check current availability before committing to an architecture that assumes in-country residency, and in the meantime document the EU-but-not-Belgian position for your data protection officer.

For Dendermonde specifically, West Europe (Netherlands) is the normal primary with France Central (Paris) as the pair. Latency from Dendermonde to that region is not usually a constraint for analytical workloads; the constraints that bite are feature availability per region and the processor chain your data protection officer will ask about.

• Belgium's federal structure means some reporting obligations sit with the regions (Flanders, Wallonia and Brussels) rather than nationally, which multiplies the datasets a single organisation must produce.
• Language legislation affects employee-facing systems: reporting used in employment contexts may need to be available in the language of the region where staff are employed.
• As the host of the EU institutions, Brussels has an unusual concentration of organisations working to EU-level reporting rules ahead of national implementation.

Where to go next

If you are weighing a platform decision for a Dendermonde organisation, a workload assessment will answer it with evidence in two to three weeks, and the written recommendation is yours whatever you decide. Book a technical review.

We work across Belgium and deliver into neighbouring markets regularly. Nearby pages worth reading if your group has sites elsewhere: Tournai, Aalst, La Louvière, Genk, Verviers and Antwerp.

Frequently asked questions