Authorization involves restricting access to resources. Hadoop provides authorization for both HDFS and all Hadoop services. In this section, we will look at how authorization can be enabled in Hadoop to secure shared resources against illegitimate access.
The HDFS authorization model is very similar to the authorization model in a POSIX system. In POSIX, each resource—files and directories—is associated with an owner user and a group. HDFS is similar to this. Permissions are given to each of these identities separately. There are separate permissions for:
The owner of the resource
The users of the group that are associated with the resource
All other users within the system
There are two permissions levels, read and write. In contrast with POSIX, there is no execute permission on files in HDFS as files are not executables. Any user, or a user belonging to a group that has the read permission r, is only allowed to read the contents of a file from HDFS...