Cloud Data Architectures: Data Lake and Data Mesh

Topic

Data Lake

Data Warehouse

Data Format

Store unstructured, semi-structured and structured data in its raw format.

Store only structured data after the transformation.

Schema

Schema-on-read: Schema is defined after data is stored.

Schema-on-write: Schema is predefined prior to when data is stored.

Usecase

• Data exploration: Unstructured data opens more possibilities for analysis and ML algorithms.
• A landing place before loading data into a data warehouse.

• Reporting: Reporting tools and dashboards prefer highly coherent data.

Data Quality

Data is in its raw format without cleaning, so data quality is not ensured.

Data is highly curated, resulting in higher data quality.

Cost

Both storage and operational costs are lower.

Storing data in the data warehouse is usually more expensive and time-consuming.