A security domain lets different scopes of the cell, such as clusters, servers, or applications, use security settings that differ from the global configuration. Not every security attribute can be overridden at the domain level, however. The attributes that can be customised include the application security settings, the user registry, the authentication mechanism attributes related to trust, the JAAS configuration, the authorisation provider, and the Trust Association Interceptor settings, which allow a domain to use its own interceptor configuration in place of the global one. Option A is correct, since the Trust Association Interceptor, TAI, is an attribute that can be overridden in a security domain.
The other attributes remain global in their scope. The SSL configuration, option C, is managed through the SSL configuration management and endpoint settings, which have their own scope hierarchy outside the domain. The LTPA authentication mechanism, option D, is a cell-wide setting, because the same keys must be shared to maintain single sign-on across the cell. The Kerberos authentication mechanism, option B, likewise is a global configuration, as it relies on one Kerberos configuration and keytab for the cell.
When using this capability, administrators should assign each domain to the intended scope, and test that interceptors act only on the applications that should be affected. Global security continues to apply to everything that has no domain assigned.
[Reference: Multiple security domains; Security attributes that can be overridden at the domain level; Trust association interceptors.]