Glossary¶
Terms used across Whiteklay product documentation.
| Term | Meaning |
|---|---|
| Air-gapped | Operation with no connection to external networks. All models and dependencies are available offline. |
| API key | A personal access token (format vda_<64_hex_chars>) used by scripts and AI agents. It is shown once at creation and stored only as a hash. |
| AST firewall | The check that parses generated or submitted SQL and blocks queries that touch denied columns, including through SELECT *, subqueries, and aliases. |
| Asset | A cataloged object such as a table, view, or dashboard. |
| Column-level lineage | Tracing how an individual column flows between tables and pipelines. |
| Datasource | A connected system (for example SQL Server, PostgreSQL, Databricks, Hive) whose metadata VDA catalogs. |
| Dataset | A table or view within a datasource. A logical dataset is a business-level grouping of assets. |
| Deny dominance | The rule that an explicit DENY always overrides an ALLOW. |
| FQN | Fully qualified name of an asset, in the form datasource.database.schema.table. |
| Great Expectations (GX) | The open-source data-quality framework VDA uses to define, generate, and run validation checks (expectations). |
| Keycloak | The identity provider used for single sign-on and role management. |
| Lineage | The upstream sources and downstream consumers of a data asset. |
| MCP (Model Context Protocol) | The standard that lets AI agents such as Claude Desktop or Cursor call VDA tools. |
| NER / PII detection | Automatic labeling of sensitive columns (for example names, phone numbers, national IDs) using a named-entity-recognition model. |
| OpenLineage | An open standard for run-time lineage events, ingested from tools such as Spark and Airflow. |
| PBAC (Policy-Based Access Control) | VDA's authorization model. Rules combine an action, a scope, and an effect (ALLOW or DENY) and are bound to roles. |
| Profiling | Statistical analysis of a table's columns: completeness, distinct values, ranges, and distributions. |
| Query Book | A collection of saved queries and their generated SQL, organized per datasource. |
| Scope | The resource a rule applies to, such as a datasource, dataset, or column (for example column:retail_banking.*.tax_id). |
| SSM (Source-to-Target Mapping) | Mapping of source columns to target columns, with transformation rules and generated SQL. |
| Text-to-SQL | Turning a plain-language business question into an executable SQL query. |