Exploring Data Transformation with Google Cloud · Topic 2 of 5
Databases, Warehouses, Lakes and Data Governance
Databases versus data warehouses versus data lakes, data governance and why openness and interoperability remove data silos, for the Cloud Digital Leader.
- 4 min
- Database: run applications and transactions. Warehouse: structured analysis. Lake: raw data of any type.
- BigQuery is Google Cloud's data warehouse; Cloud Storage is the usual data lake storage.
- Governance covers ownership, cataloging, access, quality, retention and lineage.
What it is
Organizations store data in three broad ways. A database runs day-to-day applications. A data warehouse holds cleaned, structured data for analysis and reporting. A data lake holds large amounts of raw data in any format. Data governance is the set of rules and processes that keeps that data trustworthy, secure and compliant.
The exam asks you to tell the three apart, explain why governance matters, and explain why openness prevents data silos and vendor lock-in.
How it works
Database, warehouse or lake
| Database | Data warehouse | Data lake | |
|---|---|---|---|
| Purpose | Run an application: record and update individual transactions | Analyze history across the business | Store everything raw for later use, data science and AI |
| Data | Current, operational, usually structured | Cleaned, structured, integrated from many sources | Any format: structured, semi-structured, unstructured |
| Schema | Defined before writing | Defined before loading | Applied when the data is read |
| Typical user | An application | Analysts, BI tools | Data engineers and data scientists |
| Google Cloud example | Cloud SQL, AlloyDB, Spanner | BigQuery | Cloud Storage (often queried with BigQuery) |
Warehouses and lakes are increasingly combined. With BigQuery you can query data that sits in Cloud Storage, so analysts get one view across lake and warehouse.
"Process many small transactions quickly" is a database. "Analyze years of data across the business" is a warehouse. "Keep raw data of every type cheaply" is a lake.
Data governance
Data governance decides who owns data, who can use it, how good it must be and how long it is kept. It matters because:
- Trust: decisions and AI models are only as good as the data behind them.
- Security and privacy: sensitive data must be protected and access controlled.
- Compliance: regulations set rules for storing, using and deleting personal data.
- Efficiency: people can find and reuse data instead of rebuilding it.
A typical governance process:
| Governance element | Example |
|---|---|
| Data ownership | Each dataset has a named owner responsible for it |
| Data catalog | A searchable inventory of datasets and their meaning |
| Access control | Only finance analysts can read salary data |
| Data quality rules | Required fields must not be empty |
| Retention policy | Delete customer records after the legal period |
| Lineage | Know where a figure in a report came from |
Data silos, openness and lock-in
A data silo is data trapped in one system or department that others cannot easily reach. Silos cause duplicate work, inconsistent numbers and blind spots.
Openness and interoperability remove silos:
- Open formats and standards let different tools read the same data without conversion.
- Interoperable platforms connect to many sources and clouds instead of requiring everything to move into one product.
- Avoiding vendor lock-in means an organization can change tools later without rebuilding its data.
Google Cloud supports this with open formats, connectors, and multicloud analytics such as BigQuery Omni, which analyzes data in other clouds without moving it.
A worked scenario
An insurer has claims in one database, customer emails in another system and call recordings in a file share. Each department reports different customer numbers. It lands raw files in a data lake in Cloud Storage, builds a data warehouse in BigQuery for reporting, assigns owners and access rules to each dataset, and uses open formats so other tools can read the data. The silos disappear and everyone works from one set of numbers.
How the exam asks about this
- Match the description (transactions, historical analysis, raw data of any type) to database, warehouse or lake.
- "Why is data governance important?" Trust, security, compliance and reuse.
- "How do open platforms help?" They remove silos and avoid lock-in.
What to remember for the exam
- Database: run applications and transactions. Warehouse: structured analysis. Lake: raw data of any type.
- BigQuery is Google Cloud's data warehouse; Cloud Storage is the usual data lake storage.
- Governance covers ownership, cataloging, access, quality, retention and lineage.
- Openness and interoperability break down silos and reduce lock-in.
A data lake without governance becomes a "data swamp": lots of data nobody can find or trust.
Common mistakes
Using a warehouse to run an application. Warehouses are built for analysis, not for many small updates.
Thinking governance only means security. It also covers quality, ownership and retention.
Believing a single vendor removes silos. Openness to many sources and tools is what removes them.
Check yourself
Which store keeps raw data of any type for later analysis?
A data lake.
Which Google Cloud product is the data warehouse?
BigQuery.
Why do openness and interoperability matter for data?
They let different tools and clouds use the same data, removing silos and avoiding vendor lock-in.
Topic 2 of 5 in Exploring Data Transformation with Google Cloud
Test yourself on Exploring Data Transformation with Google Cloud
The full Google Cloud Digital Leader question bank has 512 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by Google Cloud. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.