[Improvement] adding Kerberos authentication framework for catalogs

### What would you like to be improved?

As of now, there are a lot of duplicate code in the Kerberos authentication for Hadoop catalog, Paimon catalog, Hudi catalog, and Iceberg catalog.

### How should we improve?

I think we can adding a Kerberos authentication framework for code reuse and facilitating the extension of Kerberos support to other catalogs in the future.

Start by locating the existing Kerberos authentication implementations for the Hadoop, Paimon, Hudi, and Iceberg catalogs and comparing their duplicated code. Define a shared framework boundary that those catalog integrations can reuse, then verify that current authentication behavior is preserved and future catalog extensions have a clear integration path.

Written by the indexing model from the issue text.