The Few Question about Apache Atlas?

[#1]

Hello Guys,

I am using Atlas-Ranger public preview machine for exploring more on Apache Atlas but have some doubt/questions on that.

Here I have list down few of them:

1)When we create new Tag using Atlas UI,at that time atlas would also ask us to enter "attribute".what is that attribute and it's significance,why is the purpose of using attribute?

2)As all of us know that, the atlas captures metadata for all entities(like hive table,column etc.) and also allow us to assign a tag to any entity(i.e. table/column) so my question is

Where does the atlas stores those tags and metadata?

3)In the newer version, Atlas and Ranger works together I know that we can create "Tag based Policy" but the question is,

How does Atlas and Ranger works internally/what is it's internal mechanism? or

How does it stops user from selecting particular table column?

[#2 · Accepted solution]

@Manoj Dhake

The tag is a marker that Ranger tag based policies use. Before tags, you would have to create a policy that specifically mentioned the fields, tables, ect that you wanted to secure. With tags, you create a marker on that field or table or other supported entities that mark those entities as targets for the policy.

Atlas stores all of entities in manages as graph entities within a graph data base component called Titan. Titan requires a backend data store so you can choose tools like BerkleyDB, HBase, ect. Titan also needs an indexing engine to support search. You can choose from Solr or Elastic Search.

Ranger has plugins that allow it to interact with many HDP components:

• HDFS
• Yarn
• Hive
• Hbase
• Storm
• Kafka
• Knox

Using the Hive plugin, Ranger can intercept a request against a Hive table based on the schema, apply any relevant security policies, and when applicable, cause Hive to throw a security exception.

Check out this link for more details:

http://hortonworks.com/hadoop-tutorial/manage-security-policy-hive-hbase-knox-ranger/