I am just confuse with the explanation given on multiple forum for database storage for snowflake. When they say that data is stored in the form of columner and optimised files in cloud storage, does it mean S3 bucket or azure blob storage? Does Snowflake store data or uses cloud host storage ?
2 Answers
1
votes
According to the paper The Snowflake Elastic Data Warehouse (2016) - see paragraph 3.1 Data Storage:
Snowflake initially chose Amazon S3 to store table data, query results, and temp data generated by query operators (e.g. massive joins) once local disk space is exhausted, as well as for large query results. Metadata such as catalog objects, which table consists of which S3 files, statistics, locks, transaction logs, etc. is stored in a scalable, transactional key-value store, which is part of the Cloud Services layer.
Since then, and as of today, Snowflake has been made available to run also on Azure and Google Cloud.
Therefore, when setting up a Snowflake account, the user is presented with the option of a cloud storage provider to use: for AWS Snowflake will use Simple Storage Service (S3), for Azure it will use Azure Blob Storage, and for Google Cloud it will use Google Cloud Storage (GCS).
