Judy arrays are implemented as a trie and one of the cool things is that arrays can be sparse or dense. As a sparse array it is (almost) as fast as a hash table. As a dense array it basically works as a bit map which is one of, if not, the fastest data structure around.