[
https://issues.apache.org/jira/browse/PARQUET-1659?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16935553#comment-16935553
]
Xinli Shang commented on PARQUET-1659:
--------------------------------------
Yes, I understood the philosophy of designing GCM_CTR. The page size is default
1M. The footer size shouldn't matter because I commented out the code to
generate the signature for the footer.
What I suspect is that there could be some *base* overhead for invoking GCM
even you have very few data(metadata in this case) to generate the signature.
Many companies are still using Java 8 I think.
Introducing *CTR only* doesn't mean rule out *GCM/GCM_CTR*. It just adds one
more option for people to choose as a trade-off between performance and
security. Unless *CTR only* itself has a security hole, which I don't think
so, it would be nice to let people choose.
Yes, that would require to change the spec. But this is a very lightweight
change and it doesn't change any direction of the original encryption design. I
am not sure it still require PMC votes(any spec change need PMC vote?), but I
think once we have agreement on the technical part, we can bring to PMCs to
ask.
> Add AES-CTR to Parquet Encryption
> ----------------------------------
>
> Key: PARQUET-1659
> URL: https://issues.apache.org/jira/browse/PARQUET-1659
> Project: Parquet
> Issue Type: Improvement
> Components: parquet-cpp, parquet-format, parquet-mr
> Affects Versions: format-2.6.0
> Reporter: Xinli Shang
> Priority: Minor
> Labels: pull-request-available
>
> AES-GCM-CTR perform GCM encryption on metadata and CTR encryption on data.
> AES-CTR would perform CTR encryption on both.
> During Perf testing, we found AES-CTR can improve read/write performance by
> ~10% comparing with AES-GCM-CTR.
>
> I checked with Gidon and the initial assumption was that AES-GCM-CTR would
> have similar performance as AES-CTR. But with recent performance
> benchmarking, we found it is worthy to introduce AES-CTR. Since many
> companies strive for parquet performance improvement.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
