danielcweeks commented on PR #15428: URL: https://github.com/apache/iceberg/pull/15428#issuecomment-4007739084
> Does this make sense? Full control in server, with the client adapting to its choices. Polaris can choose what to sign, clients can choose various checksum options, and the only consequence is an increase in cache misses if the signing service signs more headers than before. This just feels like we're overcomplicating the actual use case and how this is beneficial to the client and the service. The only two requests that should be cached are `HEAD` and `GET`. `HEAD` requests shouldn't really have any difference in the request. `GET` requests are reused to reduce the sign request rate for range based queries when reading parquet. Each request is isolated to a single path and there's very little reuse beyond what happens within the context of a single file read action. What led to this line of exploration appears to stem from trying to implement `PUT` or other methods, which should not be cached. A server shouldn't do that because it's not secure for it to omit critical headers like checksum or size from the signature when creating/overwriting objects. That also renders those types of operations un-cacheable, so they shouldn't be considered here. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
