Bigtable (according to the paper), HBase and Hypertable all use bloom filters to improve performance of random reads. Here is some bloom filter test code in Hyp

Bigtable (according to the paper), HBase and Hypertable all use bloom filters to improve performance of random reads. Here is some bloom filter test code in Hypertable to compare Jenkins' lookup3, Hsieh's SuperFastHash and Appleby's MurmurHash2: http://github.com/vicaya/hypertable/blob/4fce96592cab3fd6d0e022a0471001b0810f9f4c/src/cc/Common/tests/bloom_filter_test.cc http://github.com/vicaya/hypertable/blob/4fce96592cab3fd6d0e... MurmurHash2 is the winner in all cases I've tested.