# Data Virtualization #dataengineering Data Virtualization is a technique for building models without moving data from sources. Data sources are connected and retained in a single location. Thanks to this, all data can be queries as if all information reside in a single space. Data Virtualization is part of [[Data Fabric]] methodologies. Main benefits are: - Faster exploration and time-to-insights - Cost savings: lower infra costs + less time spent for data management ## Layers Three main layers (bottom-up): 1. Connection Layer: interaction with data sources 2. Virtualization Layer: where virtual tables and queries are built 3. Consumer Layer: UI An additional layer can be then identified for governance catalog.  ## Sources - https://www.youtube.com/watch?v=2XB4UaBIvNI