diff --git a/sections/kerberos_the_madness.md b/sections/kerberos_the_madness.md
index 3997ab6..1af631c 100644
--- a/sections/kerberos_the_madness.md
+++ b/sections/kerberos_the_madness.md
@@ -91,8 +91,9 @@ users. When cluster node labels are used to differentiate parts of the cluster (
 more RAM, GPUs or other features), then the queues can be used to restrict access
 to specific sets of nodes.
 
-Similarly, HBase and Accumulo have their users and permissions, while Hive uses the
-permissions of the source files as its primary access control mechanism.
+Similarly, HBase and Accumulo have their users and permissions, while Hive can 
+authorize users either through its permisions model or use the
+permissions of the source files as its primary access control mechanism or.
 
 These various mechanisms are all a bit disjoint, hence the emergence of tools
 to work across the entire stack for a unified view, Apache Ranger being one example.
@@ -146,3 +147,7 @@ hence should have their replication factor increased), and which do not get
 used more then 7 days after their creation —and hence can be automatically deleted
 as part of a workflow.
 
+HBase, Hive and HDFS allow for creation and management of such audit logs.  Various 
+mechanisms are all a bit disjoint, hence as in case for Authorization, the emergence of tools 
+to work across the entire stack for a unified view, Apache Ranger being one example.
+