From email@example.com@21:1/5 to All on Fri Sep 28 12:43:36 2018
I am busy coding a database which uses perfect ordered hashing.For any two keys k1, k2 , k1 < k2 ==> hash(k1) <= hash(k2). In every trial run, 100000 randomly generated 64 bit keys are hashed to 32 bit values with zero collisions. Lots of such hash
functions can be contrived. Perfect unordered hash functions are also part of this library. This is part of a Database management system I am coding.
The indexing part of the s/w is ready and I am a bit scared of testing it and I am looking for volunteers to test it. Non-remunerative basis only. The file is not too large at all. POSIX compliant. Although it need not be, but SQL can also be the query
language for this database, but the database itself is traditional, classical ,general and fairly ideal. However the main advantage of this will be the quality of the underlying search, since it will always only scan/access very close to the bare minimum
number of records which actually satisfy the necessary condition of the query, unless the query itself is general enough to include a large number of records. From my rough calculations it's estimated capacity is close to a trillion records although more
is also possible. Note, the file storing the index will take up 276 MB as a minimum, and will grow up to 64 GB as a maximum when it contains the maximum of 4 billion entries which in the way I intend to use the index is all that is necessary for querying
the trillion records of the database system. Since records will be stored in a compressed format the database at full capacity with a trillion records in it should take up about 64 TB.