India’s Leading Domain & Web Hosting Provider
Support
Domain & Login
Blog
Microsoft Fabric & Azure Data Analytics Services
Microsoft Fabric is Microsoft's SaaS analytics platform that unifies data integration, lakehouse and warehouse engineering, real-time analytics and Power BI on a single OneLake storage layer billed by shared capacity. NTSPL designs and builds Fabric lakehouses, pipelines, semantic models and real-time dashboards for Indian enterprises consolidating fragmented reporting, with Azure Databricks where heavy Spark engineering or ML justifies it.
Microsoft Azure services in scope
The Problem OneLake and Fabric Lakehouse and Data Factory in Fabric Solve
Most Indian mid-size enterprises do not lack data; they lack one version of it. Sales figures in the ERP disagree with the CRM, plant data sits in historians that IT cannot query, and 300 Power BI reports each import their own copy of the same tables on their own refresh schedule. A typical estate has an on-premises SQL data warehouse, an Azure Synapse workspace from an earlier project, and Excel extracts mailed every morning. Each copy adds storage, refresh failures and arguments in review meetings about whose number is right.
NTSPL consolidates onto Microsoft Fabric with a medallion lakehouse in OneLake: raw data lands in bronze via Data Factory pipelines, mirroring or Eventstreams, is cleaned into silver with notebooks or Dataflows Gen2, and is modelled into gold tables that Power BI semantic models read in Direct Lake mode without re-importing. We size Fabric capacity from measured workloads, set up workspaces, domains and sensitivity labels, and keep Azure Databricks only where an existing Spark or ML investment makes it the better engine.
When Microsoft Fabric Is the Right Analytics Platform (and When Databricks or Synapse Remains)
Signals that point to data & analytics on Microsoft Azure
- More than 100 Power BI reports each import overlapping data, and refresh failures are a weekly issue.
- Leadership reviews stall on reconciling numbers between ERP, CRM and spreadsheet extracts.
- An Azure Synapse dedicated SQL pool or on-premises SQL warehouse is due for renewal or re-platforming.
- Operational dashboards need data under 5 minutes old, but nightly batch loads are the only option today.
- Power BI Premium per-capacity (P SKU) contracts are expiring and need to move to Fabric F SKUs.
- Data engineering is spread across SSIS, Azure Data Factory, Databricks and Python scripts with no single lineage view.
When a different approach is better
- If your team has a large, well-run Azure Databricks estate with Unity Catalog and MLflow, a move to Fabric is not a cost saving; keep Databricks as the engineering platform and use Fabric or Power BI only for consumption via OneLake shortcuts.
- For an OLTP application database, Fabric is the wrong tool; use Azure SQL Database or Azure Database for PostgreSQL and mirror into Fabric for analytics.
- Organizations with fewer than 20 report users and a single source system rarely need Fabric capacity; Power BI Pro with import models on Azure SQL is simpler and cheaper.
Fabric and Azure Analytics Building Blocks: OneLake, Lakehouse, Data Factory, Real-Time Intelligence, Power BI and Databricks
| Service | Role in the architecture | Choose it when |
|---|---|---|
| OneLake and Fabric Lakehouse | OneLake is the single tenant-wide data lake storing tables in open Delta Parquet format, and a Lakehouse item exposes those tables to Spark, a SQL analytics endpoint and Power BI without copies. | Multiple teams and tools each keep their own copy of the same data and you want one governed storage layer. |
| Data Factory in Fabric | Pipelines and Dataflows Gen2 ingest and transform data from ERP, SaaS and on-premises sources through the on-premises data gateway, with scheduling and monitoring inside the Fabric workspace. | You run SSIS packages or Azure Data Factory pipelines that load a warehouse and want ingestion managed alongside the data it produces. |
| Fabric mirroring | Continuously replicates operational databases such as Azure SQL Database, SQL Server, Azure Cosmos DB and Snowflake into OneLake as Delta tables in near real time without building ETL. | Reports need data from an operational database within minutes and you want to avoid hourly extract jobs hitting production. |
| Fabric Real-Time Intelligence | Eventstreams capture streaming data from IoT Hub, Event Hubs or Kafka, Eventhouses store it in KQL databases for sub-second queries, and Activator triggers actions on conditions. | Decisions depend on data less than a few minutes old, such as plant sensors, collection-centre readings or transaction monitoring. |
| Power BI (Direct Lake semantic models) | Semantic models in Direct Lake mode read Delta tables directly from OneLake, giving import-like performance without scheduled refresh copies of the data. | Large import models take hours to refresh or exceed memory limits, and reports must reflect gold tables soon after they update. |
| Azure Databricks | Managed Apache Spark and lakehouse platform with Unity Catalog, used for heavy data engineering, ML and existing Databricks codebases, which can share Delta tables with Fabric through OneLake shortcuts. | You already run substantial Databricks workloads or need advanced ML and Spark tuning beyond what Fabric notebooks offer. |
Fabric medallion lakehouse with real-time and Direct Lake serving
Fabric medallion lakehouse with real-time and Direct Lake serving
ERP, CRM and operational databases are ingested into a bronze Lakehouse in OneLake through Data Factory pipelines, mirroring and the on-premises data gateway, while streaming device data enters through Eventstreams into an Eventhouse. Notebooks and Dataflows Gen2 refine data into silver and gold Delta tables, which Power BI semantic models read in Direct Lake mode and Activator monitors for business alerts.
Select any component
Fabric medallion lakehouse with real-time and Direct Lake serving: click any of its 10 components to see its role and the design choice NTSPL makes for it, or trace the 6-step primary path.
Fabric Engineering Decisions: Capacity Sizing, Medallion Layers, Direct Lake Models and Workspace Domains
Capacity sizing and separation
Size capacity from a 2 to 4 week trial with real pipelines and reports, reading the Capacity Metrics app rather than guessing. A mid-size enterprise often starts at F32 or F64. F64 and above lets free-licensed users view Power BI content, which can remove hundreds of Pro licences. Separate engineering and reporting onto different capacities when heavy Spark jobs cause interactive throttling during business hours, rather than buying one oversized capacity.
Lakehouse versus Warehouse items
Use Lakehouses with Spark for ingestion and transformation where engineers write Python or PySpark and data includes semi-structured files. Use a Fabric Warehouse for the gold layer when the team is T-SQL-skilled, needs multi-table transactions or is migrating stored-procedure-heavy logic from SQL Server or Synapse dedicated pools. Both store Delta in OneLake, so the choice is about the engine and skills, not about where the data lives.
Direct Lake model design
Direct Lake performs well only when gold tables are shaped for it: star schemas, integer surrogate keys, V-Order enabled, and table sizes within the capacity's Direct Lake guardrails for row count and model memory. When limits are exceeded, queries fall back to DirectQuery and slow down. Pre-aggregate very large fact tables, avoid calculated columns in the model, and push logic into the gold layer.
Ingestion pattern per source
Mirroring is first choice for supported databases because it is near real time and needs no pipeline code. For SAP, use a certified connector or SAP extraction to staging, respecting SAP licensing for third-party extraction. For SaaS APIs, use pipelines with incremental watermarks. For plant and IoT data, use Eventstreams. Avoid full reloads of large tables; they consume capacity and hit source systems during business hours.
Synapse and legacy migration path
Microsoft positions Fabric as the forward path for Synapse Analytics customers, while Synapse remains supported. Migrate dedicated SQL pools to a Fabric Warehouse using the Fabric migration tooling and validate stored procedures, then repoint Power BI. Synapse Spark notebooks move to Fabric notebooks with library and path changes. Run both in parallel for one month-end close to reconcile outputs before decommissioning Synapse.
How NTSPL Stands Up a Fabric Lakehouse from Source Systems to Direct Lake Reports
Source and report inventory
Catalogue source systems, existing warehouses and the Power BI estate, identifying duplicated datasets and the top 20 reports by usage.
Capacity trial and workspace design
Run a 2 to 4 week trial on a Fabric capacity in Central India, define domains and workspaces per subject area, and set up Git integration.
Bronze ingestion
Configure mirroring, pipelines through the on-premises data gateway and Eventstreams so raw data lands in bronze Lakehouses with incremental loads.
Silver and gold engineering
Build notebooks or Dataflows Gen2 for cleansing and conformance, then star-schema gold tables with V-Order optimization.
Direct Lake semantic models and report migration
Create certified Direct Lake models per subject, repoint high-usage reports, and retire duplicated import datasets.
Governance and capacity operations
Apply sensitivity labels and endorsement, set up deployment pipelines, and review the Capacity Metrics app monthly to rebalance workloads.
What you receive
Governing OneLake: Workspace Roles, OneLake Security, Sensitivity Labels and Private Links
In Fabric the main risks are over-broad workspace access, sensitive data copied into personal workspaces or exported to Excel, and OneLake data reached over public endpoints. Controls centre on workspace roles granted to groups, OneLake security roles for table and folder access, row-level security in semantic models, sensitivity labels that persist through exports, and tenant settings that restrict who can create workspaces or share externally.
- Workspace roles assigned to Entra groups only, with Admin and Member limited to the data platform team
- OneLake security and row-level security in semantic models to restrict data by business unit or region
- Microsoft Purview sensitivity labels inherited from Lakehouse to report and exported files
- Tenant-level private links and workspace managed private endpoints for regulated data sources
- Fabric tenant settings restricting export to Excel, external sharing and workspace creation to approved groups
Fabric Capacity Economics: F SKUs, Smoothing, Pause and Resume, and Power BI Licensing
Fabric is billed by capacity units through F SKUs, from F2 upwards, either pay-as-you-go per hour or with reservation discounts, plus OneLake storage per GB. All workloads, including pipelines, Spark, warehouse queries, Eventhouse and Power BI, draw from the same capacity, with background jobs smoothed over 24 hours. Power BI Pro or Premium Per User licences are still needed for report creators, and for viewers below F64.
- Reserve Fabric capacity for steady production and use a pay-as-you-go capacity for bursty development work
- Pause non-production capacities outside business hours
- At F64 and above, remove Power BI Pro licences for view-only users
- Replace full-table reloads with mirroring or incremental pipelines to cut background capacity consumption
- Shape gold tables for Direct Lake to avoid DirectQuery fallback that consumes interactive capacity
Microsoft Fabric vs Azure Databricks vs Azure Synapse: which should an Azure customer choose in 2026?
Choose on four criteria. Team profile: business-intelligence-heavy teams with SQL and Power BI skills move fastest on Fabric, where ingestion, warehouse and reporting share one SaaS platform and capacity. Engineering depth: teams running complex Spark pipelines, MLflow and streaming at scale get more control and tuning from Azure Databricks. Existing estate: Synapse customers should plan a move to Fabric, which Microsoft positions as the successor, rather than start new projects on Synapse. Cost model: Fabric charges for reserved capacity shared by all workloads, while Databricks bills per DBU consumed on clusters you size. Many Indian enterprises land on Fabric for BI and data engineering, keeping Databricks only where ML and big Spark workloads exist, sharing Delta tables through OneLake shortcuts.
Key facts about data & analytics on Microsoft Azure
Data & Analytics on Microsoft Azure in Practice
3 data & analytics scenarios from Dairy cooperative, Automotive OEM, Hospitality: the starting position, the services chosen, the architecture, the rollout and the measured outcome.
Dairy cooperative sees milk procurement by collection centre within 15 minutes
Automotive OEM unifies dealer sales and stock data from 600 outlets
Hotel chain moves revenue management from weekly Excel to daily pickup analytics
Data & Analytics on Microsoft Azure: Frequently Asked Questions
How is Microsoft Fabric priced?
Fabric is billed by capacity, purchased as F SKUs from F2 upwards, either pay-as-you-go by the hour, which can be paused, or reserved for a discount. OneLake storage is billed separately per GB. Power BI Pro or PPU licences are needed for authors, and for viewers on capacities below F64. NTSPL sizes capacity from a trial before you commit.
We have Power BI Premium P1. What changes with Fabric?
Microsoft has been transitioning Power BI Premium per-capacity customers to Fabric F SKUs. F64 is broadly equivalent to P1 in capacity and also enables Fabric workloads such as lakehouses, pipelines and Real-Time Intelligence. Existing workspaces are reassigned to the F capacity without rebuilding reports, but you should review licensing and capacity consumption once new Fabric workloads start sharing it.
How do we migrate from Azure Synapse or an on-premises SQL warehouse to Fabric?
Synapse dedicated SQL pools and SQL Server warehouses usually move to a Fabric Warehouse, keeping T-SQL logic, with schema and data moved by Fabric migration tooling or pipelines. Spark notebooks move to Fabric notebooks. Reports are repointed to new semantic models. Run old and new in parallel through one month-end close and reconcile key figures before decommissioning.
Can Fabric data stay in India?
Yes. Fabric capacities are created in a chosen Azure region, and data in OneLake for workspaces on that capacity is stored in that region. Choose Central India or another available Indian region for capacities holding personal data covered by the DPDP Act or regulated financial data. Check the tenant's home region too, since some metadata is stored there.
How real-time can Fabric dashboards be?
With Real-Time Intelligence, events from IoT Hub, Event Hubs or Kafka flow through an Eventstream into an Eventhouse and are queryable within seconds, and Real-Time Dashboards refresh at intervals as short as a few seconds. Mirrored databases typically reflect source changes within minutes. Batch gold tables refresh on whatever schedule the pipelines run.
Do we need Azure Databricks if we adopt Fabric?
Not necessarily. Fabric notebooks run Spark and cover most data engineering for BI. Keep or add Databricks when you have large existing Databricks code, need advanced ML lifecycle tooling, or need fine control of cluster configuration for very large jobs. The two can share Delta tables via OneLake shortcuts, so this is not an either-or choice.