Final answer:
BigQuery tables use a distributed architecture and Dremel technology to provide quick querying. They are kept in columnar format. BigQuery's backend manages the complexity of data storage, sharding, and partitioning while users engage with the tables via a variety of interfaces.
Step-by-step explanation:
BigQuery tables are now kept on an infrastructure in a managed, serverless data warehouse. Data querying and storage are highly performance- and scale-optimized processes. BigQuery tables internally store their data in a columnar fashion. Because it reads only the columns required for a specific query, rather than whole rows, this format enables effective data compression and quick query execution.
Users access the tables via SQL queries using the BigQuery user interface, command-line tool, or API; they are not privy to the actual storage architecture. Since distributed storage and parallel execution environments are used in the underlying architecture, data is probably split up into several shards—physical and logical partitions that function across a global network.