Denodo: Data Virtualization Platform

Denodo is a data virtualization platform. It lets you query data from multiple sources as if it were in one place — without copying or moving the data.

What Is Data Virtualization?

Traditional approach: copy data to a central location (ETL). Virtualization approach: query data in place through a virtual layer.

Traditional (ETL):
Source A ──→ Copy ──→ Data Warehouse ←── Users
Source B ──→ Copy ──→      │
Source C ──→ Copy ──→      │

Virtualization (Denodo):
Source A ←──┐
Source B ←──┼── Denodo (virtual layer) ←── Users
Source C ←──┘

Key idea: Data stays where it is. Denodo creates a virtual view on top.

How Denodo Works

┌──────────────────────────────────┐
│          Consumers               │
│   (BI tools, Apps, APIs, SQL)    │
├──────────────────────────────────┤
│         Denodo Platform          │
│  ┌───────────────────────────┐   │
│  │  Virtual Data Layer       │   │
│  │  (views, transformations) │   │
│  └───────────────────────────┘   │
│  ┌───────────────────────────┐   │
│  │  Query Optimization       │   │
│  │  (pushdown, caching)      │   │
│  └───────────────────────────┘   │
│  ┌───────────────────────────┐   │
│  │  Connectors               │   │
│  │  (JDBC, REST, files, etc) │   │
│  └───────────────────────────┘   │
├──────────────────────────────────┤
│  Databases │ APIs │ Files │ Cloud│
└──────────────────────────────────┘

Core Steps

1. Connect — Connect to data sources (databases, APIs, files, cloud)
2. Model — Create virtual views combining data from multiple sources
3. Transform — Apply joins, filters, aggregations in the virtual layer
4. Serve — Expose unified data via SQL, REST, OData, GraphQL

Key Features

Denodo Views Hierarchy

Physical Layer    →   Base Views     (raw data from sources)
     ↓
Logical Layer     →   Derived Views  (joins, transforms)
     ↓
Business Layer    →   Final Views    (clean, business-friendly)

Example: Virtual View

-- This query joins data from PostgreSQL + Salesforce + S3
-- Data never leaves the original systems
SELECT
    c.customer_name,      -- from PostgreSQL
    s.deal_stage,         -- from Salesforce
    o.total_orders         -- from S3 (Parquet)
FROM customer c
JOIN salesforce_deals s ON c.id = s.customer_id
JOIN s3_orders o ON c.id = o.customer_id
WHERE c.region = 'EMEA';

Virtualization vs ETL

In practice: Denodo works alongside ETL, not as a replacement. Use ETL for heavy transformations, Denodo for real-time access and federation.

Common Use Cases

Denodo + Data Mesh

Denodo fits naturally with Data Mesh architecture:

Domain A (views) ──┐
Domain B (views) ──┼── Denodo ──→ Federated governance
Domain C (views) ──┘              Self-serve access

Each domain publishes virtual data products through Denodo. Consumers access them without knowing the underlying sources.

Summary