Yes, Iceberg tables great for now (interoperability).. but we need to follow 2 things in medium/long run: 1. Databricks bought company Tabular meaning they can change Iceberg with *OR* without caring to preemptively notify Snowflake. 2. Catalog wars: whoever owns catalogs and their integrations. But yes, exciting different computes sharing same storage, it's time to kill etl and data cooy/movement and all related drama.
Great feature and demo, but would had been nice to see setup in spark connectivity to Polaris catalog (some http ip address), i wonder if in snowflake side also need connectivity setup ? Maybe im misunderstanding something here but i didn't saw IPs in demo? Also the Polaris webui i suppose have it's own http IP address different from what snowflake Snowsight ?
![[LIVE] : ONE ลุมพินี 88 | คู่เอก "ป้อมเพชร vs อัสลามจอน"](http://i.ytimg.com/vi/1ZqTVVbMgbo/mqdefault.jpg)
What is unique is the possibility to write from SF and Spark to the same table. For me, that is a huge change compared to similar technologies.
Yes, Iceberg tables great for now (interoperability).. but we need to follow 2 things in medium/long run:
1. Databricks bought company Tabular meaning they can change Iceberg with *OR* without caring to preemptively notify Snowflake.
2. Catalog wars: whoever owns catalogs and their integrations.
But yes, exciting different computes sharing same storage, it's time to kill etl and data cooy/movement and all related drama.
@@emanueol eventually the catalog with best performance/price, reliable integrations and simplest usability wins.
is this frontend available in the polaris github?
Great feature and demo, but would had been nice to see setup in spark connectivity to Polaris catalog (some http ip address), i wonder if in snowflake side also need connectivity setup ? Maybe im misunderstanding something here but i didn't saw IPs in demo? Also the Polaris webui i suppose have it's own http IP address different from what snowflake Snowsight ?
How to Register a service connection, Y are u skipping it, Can you explain that part
How does it work, is the database for the REST catalog pluggable?
Will the RBAC secure the iceberg table to something not using the catalog? For example, I use spark to go after the table without Polaris.
Does Polaris have data lineage?