ranger-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Madhan Neethiraj (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (RANGER-1146) Policy engine optimization: dynamic reordering of policy evaluation order
Date Thu, 15 Sep 2016 20:08:20 GMT

    [ https://issues.apache.org/jira/browse/RANGER-1146?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15494380#comment-15494380

Madhan Neethiraj commented on RANGER-1146:

Committed to ranger-0.6: http://git-wip-us.apache.org/repos/asf/incubator-ranger/commit/efe49a88

> Policy engine optimization: dynamic reordering of policy evaluation order
> -------------------------------------------------------------------------
>                 Key: RANGER-1146
>                 URL: https://issues.apache.org/jira/browse/RANGER-1146
>             Project: Ranger
>          Issue Type: Bug
>          Components: plugins
>            Reporter: Madhan Neethiraj
>            Assignee: Abhay Kulkarni
> During initialization, the policy engine sorts the policy evaluators such that the policies
most likely to match many requests are evaluated at the top. To aid this ordering, a eval-score
is computed for each policy based on the contents of the policy - like users/groups, accessTypes,
resource-values, etc. However, this order may not be the most efficient for the access request
pattern at runtime. It will help to reorder the policies dynamically based on the number of
decisions made by each policy. For example, each time a policy determines the result of an
access request, its hit-score should be incremented. Periodically the policies should be reordered
to place policies that made most decisions at the top - so that fewer policies need to be
evaluated to determine the access request.

This message was sent by Atlassian JIRA

View raw message