Re: Extending Apache Iceberg Encryption Module

2021-03-25 Thread Gidon Gershinsky
I must say I'm impressed with the level of constructiveness and technical quality in this discussion, we're off to a good start in this project. *For POC, I think what you conclude is mostly correct, I am currently implementing the encryption spec, general encrypted file stream with KMS API, and I

Re: Extending Apache Iceberg Encryption Module

2021-03-24 Thread Ye, Jack
Sounds good, lets continue with some discussions through the doc. For POC, I think what you conclude is mostly correct, I am currently implementing the encryption spec, general encrypted file stream with KMS API, and I would expect the low level file encryption integration to take place separate

Re: Extending Apache Iceberg Encryption Module

2021-03-24 Thread Gidon Gershinsky
Sounds good, thanks. Responding to the points below: *"we can choose to store the encrypted DEKs inside the manifest or as a separated instruction file with a pointer in key_metadata, and there are tradeoffs for those approaches"* For the latter, we are running a similar mechanism in Parquet encr

Re: Extending Apache Iceberg Encryption Module

2021-03-23 Thread Jack Ye
Thanks for the feedback to the doc, we are also closely following the Parquet encryption work and would like to have that in Iceberg as soon as possible with the right architecture. Here are some brief thoughts for the points you mentioned in the email, I will add more details in the google doc:

Re: Extending Apache Iceberg Encryption Module

2021-03-23 Thread Gidon Gershinsky
Hi Jack, We're working on Parquet encryption, which is about to be released in the upcoming parquet-mr-1.12 version. Recently, we've started to look into its integration in Iceberg. It became immediately clear we need to take a wider view that covers other types of encryption in Iceberg (file stre