laskoviymishka commented on code in PR #13879:
URL: https://github.com/apache/iceberg/pull/13879#discussion_r3685080218


##########
open-api/rest-catalog-open-api.yaml:
##########
@@ -1051,6 +1051,16 @@ paths:
         table. The configuration key "token" is used to pass an access token 
to be used as a bearer token
         for table requests. Otherwise, a token may be passed using a RFC 8693 
token type as a configuration
         key. For example, "urn:ietf:params:oauth:token-type:jwt=<JWT-token>".
+
+
+        The response may include a read-restrictions field. A reader that 
supports read

Review Comment:
   > the reality is vended creds are output of GRANTs one user has on the 
catalog ... so i kind of disagree here its not identity based
   
   Yeah, fair. You're right, both are tied to the caller's identity: 
credentials come from that caller's grants, and restrictions come from the 
policy applied to them.
   
   > Read restriction are just more fine grained than the coarse grained grants.
   
   Agreed, they're on the same granularity continuum. The distinction I was 
trying to make is where enforcement happens.
   
   Vended credentials are enforced at the storage layer. The object store only 
returns files allowed by the credential, regardless of what the client does. 
Even a buggy client with a path-scoped credential cannot read outside that 
scope.
   
   Row filters and column masks cannot be enforced there. The storage layer 
deals with files and row groups, not logical rows, so the engine receives the 
raw data and has to apply the restriction itself.
   
   So the distinction is less "coarse vs. fine-grained grants" and more 
"storage-enforced vs. client-enforced." Credentials have a storage-layer 
backstop; read restrictions do not. 
   
   That's the point I was trying to make about the per-principal assumption 
carrying more weight for read restrictions.
   



-- 
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]

Reply via email to