Exposing Cassandra Bloom filter searching to the Client. Code provides Bloom filter UDF, Client side multidimensional Bloom filter (MBF) index, and server side MBF secondary index implementation.
Currently the allocation of the Buffer table is never split or joined. So small allocations may never be reused it there are not more small allocation requests. A process should be added that will
join adjacent free space blocks into a single free space block.
when a excessively large (2x request?) block is located as a free block, split the block at the end of the requested size and create a new free block.