You are viewing a plain text version of this content. The canonical link for it is here.
Posted to oak-issues@jackrabbit.apache.org by "Angela Schreiber (Jira)" <ji...@apache.org> on 2020/09/01 09:41:00 UTC
[jira] [Commented] (OAK-9189) Benchmark Results - Status Quo
[ https://issues.apache.org/jira/browse/OAK-9189?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17188312#comment-17188312 ]
Angela Schreiber commented on OAK-9189:
---------------------------------------
attached results for running the new benchmarks with the following configuration:
- Items to Read: 10000
- Number of ACEs per Principal: 1 5 10 11 21 31 (assuming that the randomly created entries are created for different paths, in which case 11 entries would hit the threshold when entries are no longer eagerly loaded)
- Number of Groups: 1 10 100 1000
despite {{Session.hasPermission}} behaving slightly different from {{AccessControlManager.hasPrivileges}} they should be somewhat comparable from a performance point of view. the results and the difference between the 2 benchmarks with 1000 principals having >5 ACEs each seems to back the observation reported. with > 10 entries for each principal the performance between the 2 benchmarks is comparable for all combinations as that those reads end up hitting the default cache which loads permission entries as they are needed.
> Benchmark Results - Status Quo
> ------------------------------
>
> Key: OAK-9189
> URL: https://issues.apache.org/jira/browse/OAK-9189
> Project: Jackrabbit Oak
> Issue Type: Technical task
> Reporter: Angela Schreiber
> Priority: Major
> Attachments: HasPermissionHasItemGetItemTest_statusquo.txt, HasPrivilegesHasItemGetItemTest_statusquo.txt
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)