Hi, Jin here from Amplitude. The real-time data store is part of a bigger columnar store we built last year called Nova (https://amplitude.com/blog/2016/05/25/nova-architecture-unde...). In designing Nova, we’ve looked at many existent solutions including Amazon Redshift and Google BigQuery, but none of them sufficiently supports all our use cases. You can read more in the linked blog.
I read that and your motivations for building nova align very well with bigquery. E.g. immutabled (big query was append only), felaxability (break out of SQL with dataflow).
Comments
Hi, Jin here from Amplitude. The real-time data store is part of a bigger columnar store we built last year called Nova (https://amplitude.com/blog/2016/05/25/nova-architecture-unde...). In designing Nova, we’ve looked at many existent solutions including Amazon Redshift and Google BigQuery, but none of them sufficiently supports all our use cases. You can read more in the linked blog.
I read that and your motivations for building nova align very well with bigquery. E.g. immutabled (big query was append only), felaxability (break out of SQL with dataflow).