You can define a group to serve as an anchor point for users who otherwise have no common access requirements. For example, engineers and scientists, as well as other problem-solving users, might have no need to access application-related data in the system. Their only interest might be in their own personal data. You can place this set of users in a single group that has no access to other data.

Also, you can define groups based on access level. For example, if PAY.DATA is a RACF-defined data set, two groups could be defined, PAYREAD and PAYUPDTE, both of which would appear in the PAY.DATA access list, but with READ and UPDATE access, respectively. Any users requiring access would be connected, as appropriate, by the group administrator.