I don’t think they would pick any SaaS liek Databricks. Companies like Meta are so big that it’s much cheaper to pay for your own solutions than any other service.
That said, most of these big companies have departments that run additional data stack other than the “central” one. I wouldn’t be surprised if they also pay for Databricks and Snowflake.
I should have been clearer, here I am keeping my question to the pure technical aspect of their lakehouse (not taking in consideration the financial aspect).
On the purely technological side what aspect of the data lakehouse couldn't be handled by the current data platform providers (databricks, snowflake)?
So if Databricks had been available back then, do you think Meta would (should) have choose their data platform?
In what way is this "modernization" not a reinvention of the wheel of Databricks solution?
Thanks for the content!
I don’t think they would pick any SaaS liek Databricks. Companies like Meta are so big that it’s much cheaper to pay for your own solutions than any other service.
That said, most of these big companies have departments that run additional data stack other than the “central” one. I wouldn’t be surprised if they also pay for Databricks and Snowflake.
I should have been clearer, here I am keeping my question to the pure technical aspect of their lakehouse (not taking in consideration the financial aspect).
On the purely technological side what aspect of the data lakehouse couldn't be handled by the current data platform providers (databricks, snowflake)?
Keeping Presto?
Thanks for the wonderful content. its great to know how meta has approached it to innovate faster.
How did you create the audio for the blog post can you please share the steps for that as well?