What is a Data Lake?

This is the defining contrast with a data warehouse: a warehouse enforces a schema on write (data must fit defined tables before it goes in), while a lake enforces schema on read (the structure is applied later, when someone queries or processes the data). That flexibility makes a lake cheap and fast to ingest into, since raw logs, JSON blobs, and files can all land in the same storage without transformation up front. The tradeoff is that a lake without governance quickly becomes a "data swamp" - enormous amounts of stored data that nobody can find, trust, or make sense of, because there's no enforced structure or catalog describing what's actually in it. Modern lake architectures address this with a metadata catalog and lineage tracking layered on top of cheap object storage. Data lakes are typically the landing zone that feeds a data warehouse or data mart after cleaning and structuring, rather than a replacement for either.