DataTerminology.comNEW
Also known as: Virtual Data Integration
Creates a unified, virtual view for querying disparate sources without moving or copying the underlying data.
DESCRIPTION
Data virtualization abstracts access to multiple sources, presenting a single logical view for users and tools. It uses connectors, pushdown optimization, caching, and security to deliver performant queries. Virtualization reduces data movement, eases governance, and accelerates integration compared to physical ETL. Typical platforms include Denodo, Red Hat, and IBM technologies. Trade-offs involve query latency, source load, and complexity of pushdown. It is useful for self-service BI, logical data fabric, and agile analytics.
ALTERNATIVE DEFINITIONS
Denodo
Data virtualization is a modern data integration approach that delivers a real-time, integrated view of data across multiple sources without physically moving the data.
Emphasizes real-time integrated views and avoiding physical movement.
Red Hat
Data virtualization is an approach to data management that allows an application to retrieve and manipulate data without requiring technical details about the data, such as how it is formatted or where it is physically located.
Highlights abstraction from format and physical location details.
IBM
Data virtualization integrates data from disparate sources, including big data systems, to create a single, unified view for users and applications.
Stresses integration across disparate sources and unified access.
RELATED TERMS
Data Federation Extract-Transform-Load Query Federation Semantic Layer Data Warehouse Data Lake Data Discovery Platform PolyBase Presto Logical Data Fabric
COMPETING TERMS
Physical Data Integration Data Replication
RELATIONS
- Broader: Virtualization
KNOWLEDGE AREAS
Data Integration & Interoperability Data Architecture Platform & Tooling Data Warehousing & BI
TARGET AUDIENCE
Data Architect Data Engineer BI Specialist Analytics Engineer Chief Data Officer
EXTERNAL SOURCES
- Wikipedia - Data virtualization
- Denodo - Vendor overview of data virtualization
- Red Hat - Concept and benefits explained
- IBM - IBM perspective on virtualization
- Microsoft - PolyBase query federation guide
- Trino - Distributed SQL federation concepts
TOP QUESTIONS
Why use Data Virtualization instead of ETL?
It avoids data copies, accelerates integration, and provides a single view with minimal movement. ETL remains useful for performance and historical storage.
Category: Alternatives
How does Data Virtualization handle performance?
Engines push down operations to sources, use caching, and optimize joins. Performance depends on source capability and network locality.
Category: Performance
What are common mistakes with Data Virtualization?
Overlooking source limits, insufficient indexing, and ignoring governance lead to slow queries and compliance risks.
Category: Mistakes
Is Data Virtualization suitable for real-time analytics?
It can serve near real-time views if sources support fast queries and caching, but high-frequency streaming may require dedicated pipelines.
Category: Opinion
How to secure Data Virtualization layers?
Implement row-level and column-level security, federated authentication, masking, and audit logging. Enforce policy centrally.
Category: Requirements
What’s a key benefit of Data Virtualization for BI teams?
A unified semantic layer simplifies access to diverse sources without tedious data movement.
Category: Example
How does Data Virtualization relate to data mesh?
It provides logical integration and access layers across domains, aligning with mesh principles of decentralized ownership.
Category: Overview
What are indicators Data Virtualization is a fit?
Many sources, strict governance, avoidance of duplication, and quick delivery timelines favor virtualization approaches.
Category: Requirements
What are risks of Data Virtualization?
Source overload, latency from remote joins, and complex pushdown logic. Monitoring and capacity planning are essential.
Category: Risks