OneLake & Medallion Architecture
Structuring your data estate into Bronze (raw ingestion), Silver (cleansed and conformed), and Gold (curated business dimensions) layers using open Delta Parquet tables, eliminating proprietary data vendor lock-in.
Microsoft Fabric Consulting Australia
Enterprise Microsoft Fabric Architecture & Migration Consulting
Ultron guides Australian enterprise data teams through pragmatic Microsoft Fabric adoption. From lakehouse Medallion architecture and OneLake shortcuts to sub-second Direct Lake semantic models and F-SKU Capacity Unit (CU) cost governance, our senior engineers help you unify data engineering, analytics, and governance without runaway cloud costs.
Platform Capabilities
Unified Lakehouse Architecture Built on OneLake
Microsoft Fabric consolidates Azure Synapse, Data Factory, and Power BI into a single software-as-a-service (SaaS) platform. We help organisations harness this unified architecture while establishing rigorous guardrails around capacity consumption and architectural governance.
Direct Lake Semantic Engineering
Bypassing traditional import refreshes and DirectQuery latency. We engineer Direct Lake semantic models that read directly from Delta Parquet into the Power BI VertiPaq engine at pure in-memory speeds without data duplication.
Fabric Capacity (F-SKU) Sizing & Governance
Configuring Fabric Capacity Units (CUs), analyzing 24-hour background smoothing curves, avoiding throttling locks, and managing compute burndown using the Microsoft Fabric Capacity Metrics app to optimize monthly Azure expenditure.
Data Factory Pipelines & PySpark Notebooks
Modernising legacy SSIS and Azure Data Factory pipelines into native Fabric Data Factory pipelines and Apache Spark notebooks, featuring automated error handling, incremental loading, and CI/CD git integration.
Engineering Focus
Enterprise Fabric Engineering Principles
Moving to Fabric requires thoughtful architecture, not just turning on a licence. We focus on performance economics, security boundaries, and operational simplicity.
01 OneLake Shortcuts & Virtualization
Rather than orchestrating expensive, slow data copies across clouds, we utilize OneLake Shortcuts to virtualize data residing in Azure Data Lake Storage Gen2, Amazon S3, and Snowflake directly into your Fabric Lakehouse, creating a zero-ETL unified semantic layer.
02 Direct Lake Fallback Prevention
Direct Lake models automatically fall back to DirectQuery when queries exceed memory thresholds or utilize unsupported DAX patterns, devastating performance. We optimize table compression, optimize V-Order indexing, and validate data types to guarantee 100% pure Direct Lake execution.
03 Purview Unified Governance
We integrate Microsoft Purview across Fabric workspaces to enforce end-to-end data lineage, automated classification of sensitive Australian tax (TFN) and health data, and domain-based access delegation compliant with Australian Privacy Principles (APPs).
Architecture Evaluation
Architectural Comparison: Traditional BI vs. Synapse vs. Microsoft Fabric
Understanding the structural differences between legacy Microsoft analytics stacks and the unified Fabric OneLake architecture is vital for accurate budgeting and strategic planning.
| Architecture Dimension | Standard Power BI (Pro / PPU) | Azure Synapse + Power BI | Microsoft Fabric OneLake Architecture |
|---|---|---|---|
| Storage & Data Format | Proprietary compressed `.pbix` VertiPaq files per dataset. | Dedicated SQL pools (relational) + ADLS Gen2 blob storage. | Unified OneLake with open Delta Parquet files and multi-cloud shortcuts. |
| Data Movement & Refresh | Nightly scheduled imports; duplicate data per workspace; memory limits. | Complex ADF ETL pipelines copying data into Synapse, then into Power BI. | Direct Lake mode: zero data movement; near real-time querying directly on Delta Lake. |
| Compute & Capacity Model | Per-user licensing ($15-$30/user/mo) or fixed P-SKU tenant capacity. | Fragmented billing across Synapse DWUs, ADF data flow execution, and Azure storage. | Unified Capacity Units (F-SKU) dynamically shared across SQL, Spark, ADF, and Power BI. |
| Developer Experience | Power Query M & DAX Desktop authoring; limited CI/CD. | Disjointed tooling across Synapse Studio, SSMS, Visual Studio, and Power BI Desktop. | SaaS web-based workspace with native Git integration, PySpark, T-SQL, and low-code Dataflows. |
| Governance & Lineage | Workspace-level permissions; lineage stops at data gateway boundaries. | Requires manual Azure Purview scanner configuration and hybrid RBAC. | Native Purview integration; end-to-end data lineage from raw file to visual chart; automated sensitivity labels. |
Decision Scenarios
When Organisations Transition to Microsoft Fabric
Fabric offers massive architectural advantages, but timing and readiness determine whether adoption produces immediate ROI or needless overhead.
- Power BI Import Limits Exceeded: When datasets surpass 10GB-100GB limits, scheduled refreshes exceed 2-hour thresholds, or overnight batch windows fail to complete before business hours.
- Data Silos Across Multiple Clouds: When operational data is scattered across Snowflake, AWS S3, and on-premises SQL servers, requiring OneLake Shortcuts to centralize analytical queries without data movement.
- Retiring Legacy Azure Synapse & SSIS: When complex, fragmented Azure infrastructure requires consolidation into a single managed SaaS platform with unified Capacity Unit billing.
- Advanced Analytics & Machine Learning: When enterprise data teams need PySpark notebooks, Delta Lake ACID transactions, and MLflow tracking co-located directly with their reporting semantic layer.
Adoption Framework
Our Four-Phase Microsoft Fabric Migration Methodology
We de-risk platform transitions by executing structured technical feasibility assessments before any production data is migrated.
1. Readiness & Sizing
Audit current Power BI workspaces, data volumes, refresh durations, and concurrent query loads to accurately estimate required Fabric F-SKU capacity.
2. Architecture & Pilot
Establish OneLake workspace hierarchy, configure lakehouses, deploy Bronze/Silver/Gold pipelines, and build a high-priority Direct Lake proof of concept.
3. Governance & Security
Implement Microsoft Purview sensitivity labels, configure row-level security on Direct Lake tables, and establish Git CI/CD deployment pipelines.
4. Migration & Handover
Cut over production reporting workloads, establish telemetry monitoring via the Fabric Capacity Metrics App, and upskill internal data engineering staff.
Analytics Ecosystem
Complete Microsoft Data Platform Consulting
Fabric integrates natively across your entire Microsoft enterprise technology environment.
Plan a Governed Microsoft Fabric Architecture
Avoid unnecessary licensing expense and architectural complexity. Speak with our principal Australian Microsoft Fabric architects to evaluate your data readiness, model workloads, and calculate accurate F-SKU capacity requirements.
Frequently Asked Questions
Microsoft Fabric Consulting for Australian Organisations FAQs
When should an Australian organisation adopt Microsoft Fabric over standard Power BI?
Organisations should adopt Microsoft Fabric when their reporting estate outgrows siloed semantic models and requires unified data engineering, lakehouse storage, cross-departmental data reuse, or advanced analytics. If your team struggles with fragile overnight refresh pipelines, data duplication across workspaces, or managing separate Azure Synapse and Data Factory infrastructure, Fabric unifies these workloads onto a single OneLake SaaS platform.
How does Direct Lake mode differ from Power BI Import and DirectQuery modes?
Direct Lake mode loads Delta Parquet tables directly from OneLake storage into the Power BI Analysis Services VertiPaq memory engine on demand. Unlike Import mode, there is zero data duplication and no scheduled refresh waiting window. Unlike DirectQuery, there is no SQL translation latency or database query overhead; queries execute at pure in-memory speeds.
How do Microsoft Fabric Capacity Units (CUs) and F-SKU pricing work?
Microsoft Fabric operates on shared Capacity Units (CUs) billed via Azure F-SKUs (from F2 up to F2048) or Microsoft 365 reservations. Fabric utilizes 24-hour smoothing for background operations (like Spark notebooks and pipeline runs) and short-term smoothing for interactive queries, preventing transient spikes from causing immediate capacity throttling. We help clients size capacities accurately to avoid over-provisioning.
How does Microsoft Fabric integrate with existing Azure Synapse, Databricks, and Data Factory investments?
Fabric supports OneLake Shortcuts, allowing organisations to instantly virtualize existing Azure Data Lake Storage Gen2, Amazon S3, and Google Cloud Storage buckets into OneLake without copying data. Existing Azure Data Factory pipelines can be migrated seamlessly to Fabric Data Factory, while Databricks Delta tables can be queried directly via Direct Lake semantic models.
Australian delivery proof
See how we deliver practical Microsoft solutions.
Review approved Australian client stories across Power Apps, Power BI, Dataverse, healthcare, mining and manufacturing.