2 ms·
We use BigQuery after trying all other options including running our own VMs with database software, 3rd party data vendors and the big managed cloud services.
by manigandham 10y ago
We use BigQuery after trying all other options including running our own VMs with database software, 3rd party data vendors and the big managed cloud services.
BigQuery's no-ops model makes a huge difference. It saves a lot of effort when all you have to do is just load up your data and run queries, no other cluster starting, sizing, tweaking or maintenance necessary.
It has real-time streaming input, nested/repeated records, powerful RegEx support and UDFs so you can run some really complex queries that you can't do with traditional SQL. It also has the cheapest storage at $0.01/gb after 90 days. They even have DML statements in beta now to support update/delete statements.
If you have lots of data to store (TBs to PBs), which can archived by time or other dimension as it gets older, but also needs to be available at any time, and have a decent amount of queries or but also need to run massive scans, and don't need sub-second response times, and want really complex query logic, BQ wins.