CompactHashtable is good for read-only hastables that are created at CDS dump time and used at run time. There are a few problems with it:

1. The insertion and lookup code is convoluted.
2. There's no clear separation between meta information (size, the buckets, etc), from the actual data.
3. They are currently stored in the MD section of the CDS archive but should really be stored in the RO section.
4. It's not easy to extend it to support new types of entries.