AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Is There a Gap in Lake House Architecture?
The lake has no servers for both data and metadata today. The way that we think so evolved with like Delta Lake or iceberg right where you stick metadata into a file. That's not going to be like performance if you compare to what let's say snowflake task which is like keep metadata in another world TV horizontally scalable old database like foundation, for example. We are trying to think of it a model where we have servers for metadata and we keep the data plane like kind of serverlesswhere in a spark jobs should be able to access the raw direct Right.